STAR COLOURS : 3
FURTHER STAR COLOUR DISCUSSIONS
a.) Colour Effects Based on Magnitude
One of the biggest influences on the colour of stars as seen in the telescope is their magnitudes. For example, in say 20cm telescope, any predominantly blue colour star of 5th magnitude will appear somewhat bluish at 9th magnitude and white (or colourless) at 13th. The reason for this is the number of photons of light detected by the eye, whose threshold in darkness will have a physical limit. This limit varies from person to person, but is certainly becomes worse with age.
For observers this means colour determination, as a rule of thumb, is about four or five magnitudes above the theoretical limit of the telescope. Based on a rough experiment conducted by me, we can adequately estimate a threshold limit in observing colour. I.e.;
m(v) = 5.42 + 3.22 × log (D (cm))
Colour Magnitude versus. Aperture :Observed, Theoretical and Average Limits of the colour perceived through telescopes or binoculars of bright field stars. Visible colour below this is unlikely due to the faintness of the star and the sensitivity of the eye to colour at low illuminations. Deep red or blue stars, or colour contrasting pairs will probably be seen by visual observers. The Table and Graphic here are only a guide!
NOTE: These limits here are only a guide. Strong colours like deep red variables and luminous blue stars probably are below this stated threshold limit. Colour contrasts for stars at the lower end of the range are likely to disappear at these particular lower magnitudes. Furthermore seeing and transparency conditions, dust or smoke, and proximity to urban skies through light pollution can drastically change the telescope magnitude limits. On the best nights, it is probably possible to get 0.5 magnitudes lower than the limit. With experience and tricks like averted vision, it might be possible to extend these lower limits.
Some double stars can be very similar in brightness and in many cases show similar colours. The easy explanation is to assume the current stellar evolution of the component stars. So stars of similar brightness are likely to be of similar ages, and therefore will appear in similar evolutionary stages for their given mass. Knowing that main sequence and smaller stars have a long life of several billion to tens of billions of years, colour — or really the observed surface temperature — will be unlikely not be too different at the time we observe them. Furthermore, as most stars have significant true separations, we can quickly discount the effects of any physical interactions between the components. Simply this means that each star has evolved independently of each other.
In the circumstances described above, the statistics show that the most common types of pairs are the yellow main sequence stars similar in nature to the Sun. Broad examples include; α Centauri, p Eridani and 61 Cygni. Another common category is then bright white or blue systems, like the southern examples of either Acrux or Beta Muscae.
Some stars also do show significant differences in magnitude, and these are more likely to have larger colour differences. Examples include Antares (α Scorpii) or Albireo (β Cygni). Such systems are likely to differ significantly by mass and evolution. As the largest stars evolve first, the primary will likely be of later spectral classes, with the companion being both earlier in evolutionary terms and hotter. This forms the common colour difference that is so attractive in the telescope. The classic example of this is Albireo, which has the K2II supergiant primary star with its hotter B8V main sequence companion. Here, the supergiant star has been first to evolve away from the main sequence, while its lower mass companion continues shines unabated on the main sequence.
Based on stellar evolution theory, in time the primary of Albireo will continue to move towards the red giant phase — increasing the contrast, before the companion star begins to evolve away from the main sequence. Probably in several tens of millions of years from now, the primary will have gone through its complete evolutionary cycle, ending as just another white dwarf (or even probably a neutron star) orbiting some bluish star — much like what we see with Sirius and its white dwarf pup star, Sirius B.
In truth, there are many exceptions to such double stars scenarios. Some systems are the reverse of this, doubtlessly caused by other evolutionary changes, such as one component entering the helium burning phase and returning only shortly as a bright star. Another distinct possibility is to undergo some mass transfer by an unresolved star or an unseen companion indivisible by the telescopic visual observer.
Since the 1920’s, there has been much discussion about the observed spectral types against the component magnitudes. This adequately highlighted in Robert Aitken’s classic “The Binary Stars” (1963) pg. 267-269. [Tables 9 and 10]. Using 9,190 double star systems with observed spectra from the ADS (Aitken Double Star Catalogue), found the following proportions;
|Spectral Type||% (percent)|
“It appears from the frequency ratios… that visual binary stars are relatively more numerous among stars of classes, B, F and G, than among the stars of other classes, the small ratios for classes K and M being specially striking… The strong contrast between the frequency ratios for class B and M… is perhaps the most striking feature… Such general statistics are of interest and have a certain degree of significance, as have also the statistical relations between magnitudes, spectral classes and colours of the components of the visual binary systems.
More specific to our discussion here, on pg. 270, Aitken also confirms what we are saying here with ;
“The relation between the colors of the components of double stars and their difference in magnitude was recognized by Struve and every observer since his time has noted that fact that when the two components are about equally bright they are almost without exception the same or nearly the same color, and that the color contrast increases with the difference in magnitude of the components, Professor Louis Bell argued that this is a subjective effect since the fainter star is generally the bluer one. Doubtless this subjective effect is often present but it is by no means the sole cause.
There are real and very striking differences in the spectral classes of the components of double stars and these correlate to their colour differences and the differences in magnitude. The absolute magnitude of the primary also enters as a factor. Thus, the astronomer H.M. von Lau in, “Über die blauen Doppelsternbegleiter (Over the Blue Double Star Companions)”; A.N.; 205, 29, (1917) [NOTE 1] and “Über die Farben der Doppelsterne” (On the Colours of Double Stars); A.N. 208, 179, (1918) [NOTE 2], respectively, also found that generally the companions to giant stars are bluer, while the companions to dwarf stars redder than their brighter primaries.
NOTE 1 : This is presented as 1918 in ADS astronomical on-line service, and the text is written in German.
NOTE 2: This is 1919 in the ADS astronomical on-line service, and is written in German.
Another interesting paper closer to modern times is also fairly easy and useful to read; See Murphy, R.E., “A Spectroscopic Investigation of Visual Binaries With B-type Primaries.” AJ.; 74, 1082 (1969)
Yet, perhaps the most important of these conclusions, is summarised adequately by Aitken;
have investigated the relations between magnitude and spectral class,
in the components of double stars, among them Dr. F. C. Leonard.”
NOTE 3: See Leonard, F.C.; “An investigation of the spectra of visual
double stars” Lick Obs. Bull.;
10, 169 (1923). See also Peter Doig’s paper; “The
Spectra of Physically Connected Pairs and the Giant and Dwarf
Theory” MNRAS., 82, 372
(1922) and G. Shajn’s (“Results of Observations of the Double
Stars and their Relation to the Giants and Dwarfs Theory”; Bull. Poulkova Obs.; 10, 276
This latter 1925 paper I was unable to obtain. It is presumably closely related to the reworked article by ; Shajn, G.; Hase, V.; “The Luminosity and the Colour of the B- and A-type Stars”; A.N., 225, 307 (1925)
Aitken also says;
“[Leonard] thorough analysis of the data for 238 pairs clearly showed“ [The other papers have conclusions that are similar to Leonard’s.];
(1) that when the two components are of equal or nearly equal magnitude, they differ little in spectral class, except in the case of a few giant stars like — γ Circini;
(2) that when the primary is a giant of spectral class F0 or later, the companion belongs to an earlier spectral class; and
(3) that when the primary is a dwarf star, or a giant star of spectral class earlier than F0, that is, when the primary is a star of the main sequence, the companion belongs to a later spectral class.
Aitken then properly concludes the discussion by saying;
“These results prove that the colour relationships observed by H.M. Lau correspond to actual differences of spectral class.”
After Aitken words here, little has really changed in our views except for the refinement to the understanding of the evolution and the processes of the formation of binary stars.
In the astronomical literature, many papers since have improved our views somewhat. Since the 1920’s, the photographic surveys were replaced with far more precise photoelectric and automated CCD imaging that are often with useful spectra. This is where newer studies of this kind have increased the sheer numbers of surveys stars. For various sources, like the Washington Double Star Catalogue (WDS), have far more data to use for similar conclusions, but the premises by Aitken for doubles still stand.
This conclusion also applies to single stars of which we have extensive data. Aitken has also properly concludes that such properties also apply to single stars, hence the relevant words by him;
Since they are also in harmony with the Mass-Luminosity Relationship (MLR) they indicate that the components of double stars are normal stars, having the same properties as ordinary single stars of corresponding mass and magnitude.
One of the greatest advances in observational astronomy was the ability to measure star colours photographically, and later by photometrically; sometimes known as filter photometry. These techniques did quite quickly eliminate the need for the quite arbitrary and unreliable selection of star colours by naked-eye visual observation. Such instrumental methods rely on taking two separate images or photometric readings using two different coloured filters of known wavelengths. I.e. Red and Blue. Any brightness variation between the two values becomes real colour difference.
Red stars will appear fainter in the blue filters, while blue stars will appear as brighter. Conversely, blue stars in red filters will become fainter, but brighter for the red stars.
One of the very first and obvious problems is the real transmission of colours was through the filters themselves. These would vary from significantly from observer to observer, whose filters would transmit differing wavelengths of light. It soon became clear that and some adopted standardised method would have to be employed to make sense of any observation. One of the old and original recommended methods was to use stable coloured chemical solutions, like blue Copper Sulphate (CuSO4) at some known concentration. This was both messy and difficult to obtain useful results — especially within wide fields or placed across the photographic plates.
|B − V||Teff (K)||Approximate|
|-4*||Pale blues||≥ −0.30||≥100,000 to 32,000||W to O8|
|-3||Pure Blue||−0.30 to −0.26||32,000 to 25,500||O8 to B0.5|
|-2||Pale Blue (Bluish)||−0.26 to −0.18||25,000 to 16,000||B0.5 to B4|
|-1||Bluish white||−0.18 to −0.08||16,000 to 12,000||B4 to B8|
|0||Pure White (White)||−0.08 to +0.11||12,000 to 9,000||B8 to A3|
|1||Yellowish white||+0.11 to +0.26||9,000 to 7,780||A3 to A8|
|2||Pale Yellow (Yellowish)||+0.26 to +0.46||7,780 to 6,470||A8 to F5|
|3||Pure Yellow||+0.46 to +0.42||6,470 to 5,860||F5 to G3|
|4||Orange Yellow||+0.42 to +0.80||5,860 to 5,690||G3 to G6|
|5||Yellow Orange (Orangish)||+0.80 to +0.86||5,680 to 5,150||G6 to K0|
|6||Pure Orange (Orange)||+0.86 to +1.08||5,150 to 4,550||K0 to K4|
|7||Reddish Orange||+1.08 to +1.34||4,550 to 3,830||K4 to K8|
|8||Orangey Red||+1.34 to +1.52||3,830 to 3,400||K8 to M2|
|9||Red Orange (Reddish)||+1.52 to +1.60||3,400 to 3,070||M2 to M5|
|10||Pure Red||+1.60 to ≤+2.00||3,070 to ≤ 2700||M5 to M10|
Historically, one of the most interesting things about seeing star colours is the wide variety of colour descriptors that have been adopted by observers. Although this has been adequately described within the first page on star colours here, there are some that may have various logical explanations.
Often we interpret that seeing white coloured stars means the observer saw no perceived colour. At first glance this might seem as trivial in difference, but the definition is something quite detailed. White stars do exist, being the moderately hot A-type spectral class stars, whose radiation shines almost equally across the visible spectrum. Possibly the visual existence of white stars often means just that the observer did not see any blues or yellows in the star in question.
If an observer does not see colour they should note this as either none seen, or at best, as colourless.
Interesting discussions often appear every now and then regarding the observation of green stars. This is especially prevalent in the observations during the 18th Century, relevant to observers like Admiral Smyth or Rev. T.W. Webb. In my view, the only star that I have really seen to display green was the companion to Antares / α Scorpii, which I thought was really closer to blue than green.
A challenging truth about the existence of green stars was likely first eliminated by M. Minnaert “Light and Colour in the Open Air.” Dover Publications (1954). If the colour green is real in stars, then I would suspect that might be caused by an optical defect of the telescope combined with effects of the eye. A similar example of this effect is what that occurs with visual observations of the planet Mars. For example, Raffaello Braga from the 33-doubles e-group (firstname.lastname@example.org) told that the;
“…companion of Pulcherrima (Epsilon Bootis) has been sometimes described as green. This is surely a colour produced by colour contrast that may be perceived when stars are observed with obstructed telescopes, as they tend to diffuse light around the primary component…”
He also reveals, that according to Flammarion, other green stars included companions to;
“…Zeta Lyr, Gamma Del and Alpha Her.”
Another “classical” green Example commonly described is often Alpha (2) Librae / α2 Lib but, I know of no other far southern ones.
The only green I see in any astronomical body are associated with several planetaries and the brightest of the emission nebulae. I.e. The Eta Carinae Nebula (NGC 3372) in Carina or the Orion Nebula, and the outer planets of the Solar System; namely, Uranus and Neptune. This is either formed by ionic emission of oxygen or [O-III] “forbidden” light or by light interactions with organic compounds like Methane.
My main disagreement with the existence of green stars is that the temperature range of these objects falls in the white A-type stars where green does not appear in the spectrum of these stars. This is because all the other colours of the spectrum equally “swamp the light” like blue, yellow and red — hence the B-V colour index value of +0.00 indicating no colour excess.
The only possible way I can think possibly to producing green stars, being likely close visual binary stars that has both blue and yellow components in the magnitude ratio of about 2:3. Then, the combined visual colour colour would be certainly be green. I know no examples of such objects.
I once did a rough search of suitable eclipsing binaries and even some ellipsoidal systems (E-II’s), and found only handfuls of examples that do met the criteria. However, all were sadly far too faint for proper visual colour assessment.
Colours for grey or ashy stars normally do mean slightly different colours, but it is still difficult to explain how the visual observer perceives such non-colours. In the purest sense, such “coloured” stars cannot exist. I personally think the observer means these stars are dull white for faint stars or even very pale yellow. Ashy stars also imply some fainter tint of red was as present.
Either white or ashy stars are likely descriptor errors or optical effects of the telescope like chromatic aberrations. Interestingly nearly all of the descriptions of these stars were in the 19th Century — and all viewed using refractors. Another point is that nearly all seem to be related to pairs that so have significant colour or display sizeable magnitude differences, suggesting these might be another form of colour contrasting effects. I cannot help feeling that there also might be some addition effects of the ambient light, which in the 19th Century was yellow candlelight. (This might be worthy experiment for some amateur to observe with both with candlelight and red light, and see if there are any colour differences. Such effects may shed some light on other problems like green, purple and indigo coloured stars. [See Below/])
The possible existence of purple coloured stars is something quite perplexing. Purple is interpreted as the combination of red and blue light — colours lying on the exact opposite ends of the visible spectrum. Stars producing such colours are simply likely due to contrasting effects.
What I believe is some observers actually mean either blue or deep blue and that the red tint is simply an aberration of the used descriptors. However, another very distinct possibility might be with visual purple or its other formal name rhodopsin, being the chemical responsible for our vision via the photoreceptor cells. Here this light sensitive compound is actually related to the monochromatic sensitive rods and not the colour cones. Studies find that rhodopsin itself will absorb strongly the wavelengths of green-blue light thus appearing more reddish-purple — the origin of the perceived colour, visual purple.
I can only think that the reasons is that such observers might have slight excesses of Vitamin A when observing, from either from their diet eaten before observing. Alternatively, the reason could even be some form of physiological effects from there surrounding observing environment. Dark adaptation may also be part of the problem, as exposure to white light tends to leave bluish tinges in the afterimage, which takes significant time — in the order of tens of minutes — before dissipating. Some sources further state that the generation of this pigment when exposed to light also takes about thirty to regenerate to normal strength after being exposed to white light. This is known as the process of bleaching, or more specifically photobleaching. Hence, there is an important established relationship to dark adaption and the need for observers to avoid white light. These real implications are important in seeing star colours after thirty minutes when exposed to colours at night so to appreciated what the observer would actually see.
The only real way of eliminating this possibility is to ensure when doing star colour estimates to have adequate dark adaptation exceed more than thirty minutes without exposure to white or coloured light. Perhaps all observers when reporting or examining star colours from such general astronomical investigations should all use this principle.
Regarding defocusing stars to see star colours it is in my view difficult to assess because it adds additional variables to telescopic assessment by the observer. Reasoning behind defocussing stars is that any stellar point is difficult to discern true colour is starlight only falls on an area with few colour cones in the eye. Enlarging the stellar point to some sizeable merged disk enables exposure to the light onto more cones, thus enabling the observer to see the underlying colour (tint).
However, the problem by enlarging the disk size also dissipates or dilutes the starlight.
If this were true, then it would furthermore also change the observed colour slightly — essentially being a colour dilution.
This is not really due to any change on the starlight but by significant differences in contrast against the background sky or even by the whole enlarged disk. I have visually noticed in some rough experiments looking at red, blue and yellow stars that the colour tint just becomes slightly lighter the larger the disk. From this point of view, I agree with other observer assessment that any enlarging of the defocusing of the telescopic image should be keep to a minimum.
For single bright stars above about 6th or 7th (though this depends on the aperture) there is enough illumination to see the colour quite clearly even though the disk was quite large. As the magnitude limit decreased, the observed tint came increasingly more washed out. A likely explanation to why this happens is that at low light levels we loose colour perception.
For more formal explanations, the deeper meaning of this likely follows this sort of logic;
Colour loss is caused by the number of photons of certain colours reaching the eye. As the light from any star is not monochromatic but a roughly offset bell curve (roughly chi-squared (χ²) in distribution and follows the famous Wein’s Law), light from the star is in fact comprises of all colours except green. What makes a star “yellow”, for instance, is that the photons that the eye is seeing contains more yellow than any other colour. If other transmitted colours exist, they can be rough adjudged by the UBV photometric system, of which, B-V, the colour-index, roughly tells us of the expected colour.
For bright stars, there are enough photons to show the distribution, so that the primary colour (tint) can be seen. For example, an orangish coloured star has more colour components distributed either side of the colour orange to contribute to the light. I.e. Yellow and orange, with all the other colours in between. The star colour is thus a colour ADDITIVE effect.
To make this clearer and simpler — if there are 1000 photons per unit of time, then there may be, say, 250 photons of yellow light, 450 orangish light and 350 of orange. The blend of all these photons is obviously an orangish coloured star.
Another star, say 350 photons of yellow light, 400 photons of orangish light and 250 of orange light will be a more yellow star, say, a yellow-orange star.
Now if the star is fainter, say 100 photons per unit time, then the colour of these various colour components will be 10% less than the brighter star. (The photons, say in the first example, would be 25 photons of yellow light, 45 orangish light and 35 of orange light.)
However, this fainter star also means that the ”sensitivity“ of the eye to the various colours comes into play. This may cause visibility to reach some colour threshold, limiting the appearance of the colour. For simplicity sake, if the colour threshold of the eye was say 30 photons per unit time, then for the bright star, this would be no difference to the colour of the star, but for the fainter one this would be more drastic. Taking our 100 photons of light in the example above finds zero photons of yellow light would be discerned, 15 of orangish light and 5 of orange. This means we see only 30% of the original 100 photons. Due to other effects like seeing, eye sensitivity, observers age, etc., the colour becomes more washed out and in fact become viewed more as a “colourless” star. I think the only way to improve and overcome this problem is with larger apertures.
The faintness of the orangish star in this case will likely be indistinguishable by an observer due to the low light levels.
By defocusing, the star artificially distributes the light over an area instead of a point. As the area increases (by the factor of π × r2 ; ‘pi times radius squared’) the number of photons per unit time is the same, but is distributed equally over the unit area. If the number of photons seen in the selected star is bright enough to seen when distributed across the “artificial disk”, then any colour should NOT be perceived below the sensitivity “threshold” of the eye. I.e. This means no observed colour.
In other words, there is an actual threshold, where stars will not show any colour, even if we do spread the light over an area. Minimising the area may brighten the stellar disk, which should increase the eye sensitivity to colour, but will be better than stellar pinpoints.
Sometimes this will not work for close double stars, whose unfocussed disks just merge as one. Of course, increasing magnification might help. Another advantage with double stars is to just compare any perceived colour differences. Even if the colours are difficult to describe, at least note differences.
One idea that I have been thinking about is in regards telescopic appearance of the Airy disk itself. What is the colour distribution in the disk and the rings themselves?
It is obvious that the rings are quite faint compared to the disk, and in most circumstances, the colour of the light in the disk is too faint to see. Of course, using an aperture stop can modify the Airy disk. There maybe also some advantage in decreasing the aperture or by increasing the general focal length — just to force the light into the rings themselves. A bit of experimentation might find some better optimum for seeing colour.
Another problem is the various magnifications of the eyepieces the observer uses, and of course with and the apparent field of view. As magnification increases, this will usually darkens the field — and here the colour changes slightly by contrast. Colours should be more obvious in darker field, so it may be worth pumping up the magnification — especially with among the brighter selected pairs.
Last Update : 29th August 2014
Southern Astronomical Delights © (2014)