| Patterns of Variable Stars | ||
|---|---|---|
Contact Binary |
Detached Binary |
|
Delta Scuti pulsating variable |
Delta Cepheid pulsating variable |
|
| Graphs are from the STARE Project Sample Results page. | ||
When do stars go dark? Is this not a silly question? The "conventional wisdom" is that stars don't go dark, at least not rapidly. As stars age and exhaust their fuel they enter a red giant stage characterized by very high luminosities relative to younger stars. The largest get even brighter when they undergo supernovas. Smaller stars after a high luminosity period as red giants, become white dwarfs that have consumed all their nuclear fuel and slowly cool over time. Variable stars, such a Cepheid variables, get brighter and dimmer in a periodic manner. There are multiple types of variable stars, some of which are close binaries, others are in the early stages of star formation and still others are nearing the end of their life [Bra01]. But none of them suddenly appear to "go dark".
Stars live for hundreds of millions to hundreds of billions of years. If the are large stars (>1.4 Msun), they will eventually explode into supernovas and leave behind black holes. These are highly noticeable events because for a brief period the power they produce outshines the galaxy in which they reside. The debris from a supernova is quite hot and fades away only over hundreds or thousands of years. Intermediate stars (G and K class), get brighter as they age by growing into red giants, then slowly fade away into white dwarfs. This process takes billions of years. The smallest stars (M class) simply fade away, a process that takes hundreds of billions of years, which is much longer than the universe has existed. So stars don't suddenly go dark. Or do they?
One question to ask would be what would happen if an astronomer observed a star going dark. To deal with the large numbers of stars in the fields of photographic, or now electronic, images, astronomers use devices called "blink comparators" to compare two identical star fields to rapidly identify stars whose brightness is changing. This approach identifies the quasars, supernovas, and variable stars that are later studied more intensively. The astronomers looking for quasars and supernovas are looking exclusively for images becoming significantly brighter. The astronomers studying variable stars are looking for moderate changes in brightness.
There is a class of stars, actually binary stars, that show a rapid change in brightness. These are situations where a white dwarf, neutron star or black hole is in very tight orbit around a normal star or red giant and is accumulating matter from this star. The high gravity of the neutron star enables the ignition of the new nuclear fuel producing a hydrogen or helium "flash" that produces a significant change in brightness [Str2000]. Once the fuel is consumed the star returns to its former brightness. This process may take months to years.
Astronomers do not expect stars to go dark, so they are unlikely to notice them. They are biased towards objects "appearing". If they want to study something, be it a supernova, an asteroid, etc., they generally have to "see" it to do that. They notice things appearing, that are highly unlikely to be stars [Zhu97]. If they discover a star disappearing, they would most likely write it off to problems with the camera, film or CCD, data transfer errors, or occultations by objects in the path such as planes or orbiting satellites. Another possible source for stars disappearing is nearby objects such as asteroids or the dark burnt out remains of comets that orbit into the line of sight between the astronomer and the star. The final possibility for a star going dark is a variable star dimming below the detection limit of the telescope and its detectors (film or CCDs). Most astronomers are unlikely to have the time or interest to investigate simple instances of "stars" disappearing unless they are surveying objects within our solar system and expect such phenomena to occur [Sto1998, Stu2001, NEAT, Coo2000, Mul2000]. In those cases the object has to reappear to receive further study. For example, the CTIO telescope has been involved in a distant supernova survey [Ham1993] that requires repeated imaging of a large section of sky. The discovery of a Kuiper Belt object in their data was not made by the astronomers conducting the survey, but by 3 high school students in an "Introductory Astronomy" class [Sin01]. Further study may require an instance of a star "disappearing" in one location and "appearing" in another location to allow detailed investigation, e.g.
"The subtraction wasn't perfect: It left mostly a grayish background with a few obvious artifacts. However, the subtraction image was good enough so that the students were able to spot the unusual pair of dots. One dot was black, suggesting that an object on the first image had disappeared from the second. And a white dot was visible about 1 centimeter from the black dot, indicating the same object had merely changed position."Without such clear examples, astronomers could spend many fruitless hours of telescope time searching neighboring regions of space in search of "disappearing" stars. Even those astronomers studying variable stars require that the star that "disappeared" must become visible again so it can be classified as a "true" variable star. Those instances where a star when dark and stayed dark would be written off to infrequent objects such as asteroids or comets that simply moved to another location in the sky. An example of the "glitches" encountered by astronomical studies may be found in [Hoo1993]:
"They devised a simple blink-like search method which quickly produced a very blue object not on published star charts. Their elation soared with the realization that they had discovered an `exotic variable blue object.' However, this marvel soon disappeared from the photograph, a tenacious dust particle which finally gave way. The disappointment was deep but short-lived, and they focused on the enjoyment and challenge of the process."Blink methods have served as the basis for many of the historical discoveries in astronomy. One can hardly fault astronomers who had to spend thousands of hours peering through these devices for not paying much attention if, from time to time, they encountered stars that "disappeared". Today we are in a better position because the computers may do the image comparison. However few facilities before now have had the data capacity to store large numbers of images that would allow the study of stellar magnitudes over longer time scales.
We are now the start of an era when we can legitimately consider the properties of and search for "stars going dark". What kind of stars would do this? Presumably those stars that have habitable planets on which are living technological civilizations that require additional power to supply the needs of their growing civilization. The standard of living of humans is tied fairly closely to their personal power consumption. If you have more power available, you have a higher standard of living. The production increasing numbers of power consuming devices that can do useful work, such as computers or even robots, means that a civilization can utilize ever increasing amounts of power. As was pointed out by Freeman Dyson in 1960 [Dys1960], even at the low growth rate of 1%/year, civilizations achieve a consumption level equivalent to the power output of their star in only 3000 years.
In advanced civilizations there are number of uses for power:
One central concept considered in the future evolution of humanity has been the colonization of space by building human habitats there [ONe1974, ONe1976]. This resulted in the idea that civilizations like our own would construct habitats that must be maintained at 300°C and that insufficient material was available to enshroud a star [Pap1985]. Thus, the searches for advanced civilizations that have been conducted to date have focused on locating stars with incomplete Dyson shells. Jun Jugaku has looked for stars that show a slight increase in infrared emission [Jug1990-99] and failed to find any indications of these incomplete Dyson shells. However, there is no reason to believe that the ability to harvest the greatest amount of available energy would be sacrificed to the maximization of living space. Solar collectors can be extremely thin, and the mass of small planets can be used to produce star enveloping satellites that would capture all of the energy. Similarly the entire power output of a star can be consumed by nanocomputers [Dre1992] with a mass of only only 3×1016 kg of material (0.000042% of the moon's mass). So we must face the possibility, that advanced technological civilizations will make their stars "go dark".
What would this look like? As the star is enveloped with solar arrays, there should be a decline in brightness. This may be an exponential decline if, as power becomes available, it is used primarily to accelerate the conversion of asteroidal or planetary material into solar collectors. The energy does not however disappear, it is instead gradually shifted into the infrared, because the heat produced by the activities of a civilization must be radiated away. Astronomy looks at stars at different wavelengths from the ultraviolet thru the infrared in bands known as UBVRIJKLMNQXZ. These correspond to Ultraviolet, Blue, Visible (green/yellow/orange), Red, Infrared, with the remainder covering near, mid and far-infrared bands. Infrared astronomy is very limited on the surface of the earth due to the absorption of infrared radiation by molecules in the atmosphere, particularly water vapor. Additional difficulties arise due to the stage of development of infrared detectors and the problem of cooling the optics of a telescope. To see the process of stars going dark, it would be necessary to observe a decrease in brightness in the UBVR bands with an increase in the IJK... bands. Dyson and Jugaku have assumed civilizations would radiate their radiation at liquid water temperatures (200-300°C) corresponding to a peak at 10-microns (the N-band). This band is rarely used by astronomers on earth due to background noise from the telescope optics. Instead, searches are done for more moderate increases in the IJK bands.
We are now entering an era however, where survey studies may provide sufficient information to study thousands to millions of stars for the possible construction of enveloping solar arrays by technological civilizations. If astronomers notice a star whose visible brightness is decreasing, that is the first hint that a star should be examined more closely. As a civilization with an optimally positioned planet (e.g. Mercury), with robust self-replicating manufacturing capabilities could enshroud the star very quickly, possibly within a few days. More likely the process will take longer because the material may need to be moved over large interplanetary distances and the development of robust self-replicating machinery may require significant development time. Further, if our solar system is an example, carbon seems to be the material most likely to be used for self-replicating systems (either natural or of the nanorobotic variety) and most of the carbon is in the outer, colder regions of the solar system in the form of CO2, where there is little power initially available to collect it and relocate it to more useful locations. It is not unreasonable to expect that years to decades would be the expected time frame for a decline in brightness. Studies have been done that show cases like these do exist [Bon1995], though these are currently believed to be due to long-period variable stars.
If we discover a star whose brightness is decreasing, the second indicator would be a shift in the temperature (or color). Astronomers typically look at the difference in magnitude between adjacent bands to determine this information. If the B-V value is positive the star is very hot. If the B-V value is negative, the star is cooler. The same can be applied to V-R, R-I, etc. So, the hallmark of a star enshrouding process is for the B-V value to become increasingly negative, while the V-R value becomes increasingly positive over time. This would continue through the other bands.
Astronomy has now entered the survey era. Large ground based surveys such as the 2MASS, DENIS, and SDSS surveys are underway. The 2MASS survey, using the JHKs bands is covering both the Northern and Southern Hemispheres and is largely complete. Its 2nd data release has been published (see Explanatory Supplement). The point-source catalog lists 162,213,354 objects. The DENIS survey, using the GJKs bands is operating in the Southern Hemisphere and is still underway. The published data includes 17,471,214 point sources. The SDSS is underway and will cover much of the Northern Hemisphere in a unique set of bands designated u',g',r',i',z', covering approximately the range from U-J (documentation here). The ground based surveys help to a limited degree, because they do not generally plan to resurvey areas that have previously imaged, so there is no data on stellar variability over time. To get information regarding variability over time would require merging the data from two of these surveys. This would entail the formidable problems of reconciling the different wavelengths they use. Having only two data points would be problematic, so additional information would be required from other catalogs or digitizations of plates from the photographic astronomy era.
It would be valuable if survey missions, such as those looking for supernovas, could be used to compare identical fields on a regular basis, not for things getting brighter, but instead for things getting dimmer.
Past and planned astrometric missions do seem provide the possibility
for a body of data from which potential candidates for stars undergoing
enshroudment may be extracted. Table 1 lists
these stars and the current or expected number of stars they will study.
| Table 1 | |||
|---|---|---|---|
| Mission | Era | # Stars | Mag. Limit |
| Hipparcos |
|
118,218
|
|
| Tycho-2 (from Hipparcos Star Mapper) |
|
2,500,000
|
|
| DIVA |
|
35,000,000
|
|
| FAME1 |
|
40,000,000
|
|
| SIM |
|
20,000
|
|
| GAIA |
|
1,000,000,000
|
|
The Hipparcos
catalog contains 118,218 stars, of which 2,712 (2.2%) are periodic (solved)
variables and 5,542 (4.6%) are unsolved variables (i.e. those for which
a period cannot be determined). Table 2 shows
the counts and fractions of these stars that can support life (classes
F2-M3) after removing those classified as giants, dwarfs or part of multi-star
systems.
We may conclude that ~1000 (1.2%) of the stars in the Hipparcos catalog
have both the class and variability that warrant further study. Applying
these ratios to the future studies and we can see that a large number of
candidate stars will be available for consideration as candidates for study
as stars that may be going dark. It is worth noting that when a selected
set of variable stars is examined, that long-period variables that decrease
in brightness are listed as "novae returning to quiescence light" [Zeb1998].
Such stars constituted 43 of 116 (37%) stars from the OGLE-I database.
That number seems rather high when entire galaxies only experience a few
dozen novas per year [IaGRA].
| Patterns of Variable Stars | ||
|---|---|---|
|
Contact Binary |
Detached Binary |
|
|
Delta Scuti pulsating variable |
Delta Cepheid pulsating variable |
|
| Graphs are from the STARE Project Sample Results page. | ||
| Star Counts Resulting from Various Star Surveys | ||||||
|---|---|---|---|---|---|---|
| ROTSE I | ASAS I | STARE | Tycho I | Hipparcos | GCVS | |
|
Ref:
|
|
|||||
| Area |
|
|
|
|
|
|
| Time |
|
|
|
|
|
|
| Number of Stars |
|
|
|
|
|
|
| Filters |
|
|
|
|
|
|
| Variable Types |
|
|
|
|||
|
|
|
|
||||
|
|
|
|
||||
|
|
|
|||||
|
|
|
|
||||
|
|
|
|
||||
|
|
|
|
||||
|
|
|
|||||
|
|
|
|||||
| Total: |
|
|||||