Tuesday October 4
The atmosphere isn't transparent at all wavelengths. It transmits radiation in the optical and radio bands of the spectrum. These are the only regions that we can study very well from the ground. This is a good thing from a biological perspective, as short-wavelength radiation (uv, x-ray and gamma ray) is very hard on organisms. The atmosphere is opaque to most of the IR part of the spectrum also. This is because the atmosphere actually radiates at IR wavelengths (heat). The atmosphere is opaque to long-wavelength radio waves because the ionosphere acts like a mirror for that range in wavelenths, and thus reflects the radiation back into space. Or from the ground back down to the ground.
Because of our biological bias, and the physics of the atmosphere, the first three hundred years of telescopic astronomy was devoted to optical telescopes. Given that, I will focus most of my discussion on them. The basic ideas are similar in other wavelength regimes. But the details differ.
Telescopes do two basic things:
Of these, light-gathering power is the more important. The light gathering power of a telescope varies directly with the area of its primary mirror or lens. This follows because for astronomical purposes, we observe objects that are so distant that the light from them can be described as a set of parallel waves (the curvature of the wavefront is negligible).
Consider the comparison of a human eye (with a pupil diameter of about 5mm) with the Andreas Observatory 0.5m. The ratio of their areas is 10e4. This is the same as the ratio of the fluxes they collect from a star. Expressed in magnitudes, this means one should be able to see stars 10 magnitudes fainter through the 0.5m than one can with the naked eye. Given a nominal limit of 6th magnitude for naked eye stars, that means an 0.5m telescope should allow visual observation of 16th magnitude stars.
Carrying this further, the eye "integrates" for approximately 0.05 sec. Thus if we integrate on a source for 500 seconds we should get another factor of 10e4 in limiting flux, or another 10 magnitudes. This implies that a 500 second observation with Andreas should reach a limiting magnitude of 26th.
These estimates are overly optimistic. The sky here is neither dark nor clear enough to reliably reach 6th magnitude. Also, that is the depth attainable if one has excellent eye-sight. The jump to 16th magnitude assumed no loss in any of the optical elements of the telescope. The further jump to 26th magnitude assumed no losses due to detector inefficiency. I'll guess that the real limit at Andreas for a 500 second exposure is between 22nd and 24th magnitude.
The limiting resolution possible in any optical system is imposed by diffraction. Most of the power in an ideal diffraction pattern is in the central maximum, which has an angular extent of
This ignores a constant term that depends on the shape of the aperture. For a circular aperture, the exact result is
Note that one must measure theta in radians!
We can use this relationship and the example of the human eye again. Now we specify a wavelength of 550 nm (the middle of the optical band). This gives us a diffraction limit of about 23" for our eyes.
If one wants a diffraction limit of 1", one then needs about a 10cm telescope. One sees, further, that the diffraction limit of something like the Andreas 0.5m is only 0."2, and that of (say) a 10m-class telescope is 0."01. Such resolutions are, however, not practically attainable because of atmospheric effects.
A basic result of transmission optics, known as Snell's Law, states that light is refracted (the direction of the light ray is deflected) when light passes from one material to another if those materials have different indices of refraction. The index of refraction, n, is equal to 1 for a vacuum, and greater than one for all normal materials. When light passes from a lower to a higher n medium, it is deflected toward the perpendicular.
The index of refraction of air is a function of density, such than n increases with increasing density. Therefore light from a star (not on the zenith) is progressively deflected as it passes through the atmosphere. One consequence of this is that the observed positions of stars are not quite their true positions. This effect grows larger as the line of sight approaches the horizon.
A second consequence of this effect results from atmospheric turbulence. The path that light from a given star takes through the atmosphere does not have uniform density stratification with time. It changes quite rapidly due to turbulence in the atmosphere. Thus the amount of atmospheric refraction changes on short time-scales. This causes the apparent position of a star to dance around, and results in a smeared-out image when integrations are taken for longer than about 0.1 second.
This atmospheric seeing limits the angular resolution of a ground-based telescope. The seeing varies from site to site, and from night to night (or even hour to hour) at a given site. A "good" site will have seeing of around 1". The best sites have seeing of around 0."3. At Andreas the seeing is very rarely better than 2".
The consequence of this is that large ground-based telescopes do not deliver images that are anything close to their diffraction limits (there are ways to improve on this that we shall return to shortly).
Consider the case of a simple "camera" with a single lens, and a detector in the focal plane. If two stars have some angular seperation of theta, they will form images on the focal plane that are seperated by some distance x. If one considers the geometry, and examines situations in which the small angle approximation is valid, one can see that the linear magnification of such a system is simply
Here f, the focal length, is the distance between the lens and the focal plane. Thus the larger the focal length, the more image magnification we get. But the total flux is fixed, so the larger the magnification, the less light per unit area strikes the detector. The image brightness fall as f^-2 (and is also proportional to D^2). This leads to the definition of the focal ratio:
Notice that this imposes a set of trades offs on the optical system. One can "stop down" the lens to produce sharper images, but at the cost of lost light.
Real optical systems have imperfections that result in image distortions. These imperfections can be catagorized as either axisymmetric (spherical aberration, for instance) and non-axisymmetric (astigmatism). All optical systems have these problems as some level. But there is another aberration that is intrinsic to the physics of lenses. That is chromatic aberration. Chromatic abberation occurs because the index of refraction is a function of wavelength. Essentially, a lens bends a beam of blue light more strongly than a beam of red light. This means that a given lens will have a shorter focal length for blue light than for red light. Thus, if one places a detector at (say) the blue focus, then the red light is out of focus. This problem can be somewhat alleviated by use of lenses called achromatic correctors. Such a lens reduces but cannot eliminate the problem.
Telescopes come in two basic flavors:
It is typical to use refractors to discuss basic telescope optics because the ray-tracing is easier to follow. The issue that drives the optical design is the need to have the distance between the objective (primary lens) and the eyepiece equal to the sum of their focal lengths. The resulting angular magnification of the system is then the ratio of the two focal lengths.
Refractors were the original telescopes, invented around 1606. But the last large, research-grade refractor was built more than a century ago now. This was the Yerkes Observatory 40" in Wisconsin. There are several reasons why refractors have been largely abandoned for research purposes. The first of these is chromatic aberration, discussed above. Notice how long and thin the 40" is. This is, in part, a means of lessening the chromatic aberration problem. If the focal length of a lens is very long, the relative amount of chromatic aberration is less than it is for a short focal-length lens. Also, recall above that one gets more linear magnification with long focal lengths.
The second reason is that one has to have two excellent optical surfaces AND a good optical interior for a lens, but only a single good surface for a mirror.
The third reason is that one must mount the lens at the top of the telescope, and steer from the bottom. Whereas with a reflector, the mass of the mirror is a the bottom of the telescope. The mechanics is simply much easier. Note that the factors discussed above that drive refractors to have long focal lengths make this mechanical problem even worse.
Reflectors have mirror-based optics, and so do not suffer from chromatic aberration. A mirror needs only have one good optical surface, instead of two plus good interior quality. And the mechanics of putting the mass near the ground are simply much easier to support.
I have been tossing the phase "good surface" about. In this context a "good" surface is one that does not introduce image distortions that are significant compared to the diffraction limit of the telescope. In round numbers, this is about 1/20 of the wavelength of light in question. For optical telescopes this amounts to having the figure exact to within about 25 nanometers.
One difficulty with reflectors is that the light is focussed on the same side of the mirror that it comes in from. This means that some part of the optical path has to be blocked by a secondary mirror or an instrument package. As an example, the old Prime Focus cage at the Palomar 200" blocked about 1m of the mirror. People observed from inside the Prime Focus cage, and this imposes a size limit of about 1m on such a facility. For a 5m primary, that only blocks about 4% of the light, but on a substantially smaller telescope, the loss would be prohibitive.
There are a number of designs for reflectors. If the telescope is a small one, to be set up in the driveway, and used with eyepieces, one typically wants a Newtonian system, with a secondary mirror at the top to direct the light to an eyepiece. (Newton invented the reflecting telescope). If one is designing a large telescope, one does not want to use Newton's design, as this puts the instruments high up in the air, and mounted off-axis. Instead, one could put them a prime focus (which is simple to design, and has the instrument package on-axis), or one can put a secondary mirror at the top to direct the light back down through a hole in the primary, and to an instrument package underneath the telescope. Such a design is called a Cassegrain. This has the benefit of keeping the (potentially very massive) instrument package near the ground.
As noted earlier, the main driver behind the desire to build larger telescopes is that they collect more light in less time than smaller telescopes. The Yerkes 40" was the largest telescope in the world when it was built in 1897. It was succeded by the DAO 72" (1915), the Mt. Wilson 100" (1919), and the Palomar 200" (1940). No one built a telescope substantially larger than the 200" for better than 40 years after that (note that WWII caused the commisioning of the 200" to be delayed until nearly 1950). There are a number of technological reasons why this is so.
The first of these has to do with Telescope Mount Design. The two basic styles of telescope mount are the Polar Mount and the Altitude-Azimuth Mount. Polar mounts are aligned along a polar axis, and thus need only turn in one direction. Further, in order to track, they simply have to turn at a constant rate, equal and opposite to the Earth's rate of rotation. But they are very unweildy to build, and impose real limits on the size of the telescope that one can afford to construct. Alt-Az mounts are much more mechanically forgiving, as mass can be loaded symetrically. But steering them requires variable rotation about both the altitude and the azimuth axes. So one needs sophisticated computer control to steer Alt-Az telescopes. As soon as this became available (in the late 1970s), that became the design of choice. The consequence of this was not that people began building larger telescopes, but rather that 4-5m class telescopes became much more affordable. In 1970 there were three such telescopes in the world, and by 1980 there were at least a dozen.
Thursday October 6
The development of computer-control adequate to point and guide large Alt-Az telescopes did not, by itself, result in the construction of telescopes larger than the 5m-class. The reason for this is the mass of a traditional mirror scales as the cube of its radius. Thus a 10-m traditional mirror will be nearly an order of magnitude more massive than the Palomar mirror. Supporting such a mirror is prohibitively expensive.
Beginning in the late 1970s, there have been several succesful approaches to building large mirrors:
Seeing is not all due to upper-atmospheric turbulence. In fact, a great deal of the image distortion in older telescopes is due to tube and dome seeing. Because of this, a large amount of effort over the last few decades has been invested in making design changes to telescopes and to domes to improve the stability of the air-flow through them, and to lessen the thermal time constants of mirrors that can drive surface convection.
Efforts to account for atmospheric turbulence have also led to dramatic improvement in the image quality from large telescopes at good sites. Two currently developing techniques are working to overcome this limit imposed by the atmosphere, and deliver true diffraction limited optics to large ground-based telescopes. These techniques are called Active Optics and Adaptive Optics.
In both of these types of systems, one monitors a bright star near one's target of interest, and uses wavefront sensors to monitor the change in the shape of the image of a point-source (the "point-spread function") on timescales of 0.1 second or so. An Active Optics system is one that changes the tip and tilt of the secondary mirror in response to the changing turbulence in the atmosphere. Such systems are becoming standard on large telescopes. An Adaptive Optics system is one in which the actual shape of the secondary mirror is altered by actuators (again on timescales of 0.1 second or so). This is a much more demanding technology, and is really only at the prototype stage.
A good site should have the following properties:
As one can see from the list above, the qualities of a good site are pretty demanding. Thus when such a site is found, it tends to attract a large number of telescopes. Mauna Kea is such a location, as are a number of sites in Andes in northern Chile. Antarctica is also a tremendous site, especially for IR work. This is so because the air is so cold that nearly all the water is frozen out. Water-vapor is one of the biggest sources of opacity in the 10-30 micron range. Thus the south pole is one of the best mid-IR sites on the surface of the Earth.
Of course, if one goes into space, then all the issues with the atmosphere are avoided. One gets diffraction-limited imaging "for free" (not counting the cost of the spacecraft and launch). And one can observe at all wavelengths.
There are two basic sorts of instruments that one puts onto a telescope.
A spectrograph is a tool to disperse the light of an object into a wavelength map. There is an enormous amount of information in the resulting spectrum .
Astronomers began using photographic emulsions to record images as soon as photography was invented (1850s). Photographic plates continued to be the main means of storing images until the 1980s. The change from photographic techniques was brought about by the development of Charge-Coupled Devices (or CCDs). These are the devices in you cell-phone cameras.
CCDs have three major advantages over photographic emulsions.
Observing with a CCD, or with any photon-counting detector, one is dealing with a system in which the Signal to Noise of an observation scales roughly as the square root of the number of counts. That is, the statistical significance of an observation follows Poisson statistics (to first order, anyway -- we deal with lots more of the ugly details in the 300-level observing classes). A summary of some of the calibration issues can be found in my notes for my Spectroscopy course.
If one shines white light through a prism, the light is dispersed into a spectrum. However, the spectral resolution of a prism is low. Thus most astronomical spectrographs use diffraction gratings to disperse the light. A diffraction grating is a reflective surface with parallele lines cut into it. If these lines have a seperation d, then they produce a spectrum by dispersing the incident light according to the following relation:
Here the m is a positive integer, called the order.
One typically speaks of the Resolving Power or Resolution of a spectrograph. This is defined as the ratio of the wavelength of interest to the difference in wavelength that is just barely measurable. For a diffraction grating, with a line density of N, the resolution goes as follows:
Very densely grooved gratings can have R up to 10^5 or so. But most general purpose spectrographs have resolutions in the range of 1000s.
One can construct a filter that blocks out light from all but a fairly narrow slice of the spectrum by using layers of material and taking advantage of interference effects. Such filters are quite pricey, but they allow things like imaging of extended sources in the light of particular emission lines.
As UV photons have smaller wavelengths than do optical photons, a given sized mirror will have a smaller diffraction limit (GOOD!), but the tolerance of the mirror figure will be correspondingly higher (BAD!). You can't have it both ways.
The main difficulty with UV observations is that the atmosphere is essentially opaque at wavelengths shorter than about 3500 Angstroms. Thus, except for very near-UV work, all UV astronomy must be done from space. There have been quite a number of UV astronomy satellites over the last 30 years. Because of the figure-tolerance issue, most of these have been purely spectroscopic missions (Copernicus, IUE, EUVE, FUSE). Three exceptions to this are the UV-sensitive cameras on HST, the ASTRO-1 and 2 deployments of the Ultraviolet Imaging Telescope from the Space Shuttle, and the current GALEX satellite. The UIT used photomultipliers to down-convert UV photons to optical, and then to expose photographic film. The resulting images have rather poor spatial resolution (several arc-seconds, but the field of view is 10-15 arcminutes). By comparison, the UV cameras on HST have ~0."1 resolution. But the field of view is only 2'.
Here we are dealing with wavelengths longer than in the optical, so the trade-off between the diffraction limit and figure tolerance is reversed from the UV case. In particular, the diffraction limit of a 1m telescope, when observing at 10 microns is about 2" (worse than nominal atmospheric seeing). So mirror size starts to matter from an image-quality standpoint.
Another problem in the IR is that we are sitting in a 300K heatbath, and so everything radiates at 10 microns. One must have very good, and very accurate temperature control not to be swamped by background in the IR. This leads to a different observing strategy than in the optical. In the optical, one might take a few long-exposure images of a given target to reach the desired sensitivity. In the IR, the detectors will saturate in a minute or less, so one must take a large number of short integrations, and stack them up afterward.
The atmospheric opacity in the IR is driven by line absorption, and is dominated by water vapor and OH. Thus altitude is even more important in the IR and the optical. At an elevation of 4km or so, one is already above most of the atmospheric water vapor. For wavelengths longer than 30 microns or so, this is still insufficient. The solution is to fly airplanes into the stratosphere, and install telescopes on them. The Kuiper Airborne Observatory was such a flying 0.9m telescope. It operated from the 1970s until the mid-90s, when it was decommissioned to devote resources to the succesor SOFIA. SOFIA is a 2.6m telescope being installed in a 747. It should begin routine operations in the next year or so.
If the stratosphere isn't high enough, one must resort to scientific ballooning or to satellites. Satellite work in the IR goes back to the pioneering IRAS mission in the mid-1980s. IRAS surveyed the sky in four bands at 12, 25, 60 and 100 microns. It provided the first survey at these wavelengths, and thus led to huge advances in our understanding of the energy balance in the interstellar medium. In the 1990s, the Europeans flew the ISO mission. ISO did not do a large scale survey, but it included a 200 micron channel, and it had spectrscopic capability. The current major IR mission is Spitzer/SIRTF. The telescope originally provided diffraction limited images from 3-160 microns, and is thus leading to almost as big an advance as IRAS did 20 years ago. The liquid He coolant on Spitzer ran out about a year ago, so now the only functioning instruments are the 3.5 and 4.5 micron cameras.
Radio astronomy is the only branch of astronomy to begin by accident. It goes back to work by Karl Jansky in the 1930s. Jansky was a radio engineer, and was working on trans-Atlantic radio communication. In particular, he was trying to identify noise sources. And he noticed that one part of the background noise would get louder once a day. With further data, he realized that the source of this noise increase kept Sidereal time. After a few years of work, he realized he had detected radio noise from the Galaxy.
The professional astronomical community did not pick up on this. Instead, the next player was an amateur astronomer (and professional radio technician) named Grote Reber. Reber built himself a backyard radio telescope, and mapped the sky with it. He even got his results published in the Astrophysical Journal. Still, the professional astronomical community showed little interest. This is because of the poor angular resolution imposed by radio wavelengths. As an example, a radio dish with an aperture of 30m, observing at a wavelength of 21cm will give an angular resolution of
Some 2000 times worse than the 1" optical seeing. This meant that it was impossible to match Reber's radio sources with any known optical sources. So mostly, the pros just shrugged, and went back to their optical work.
The real boom in radio astronomy began after WWII. This is mainly because there were a lot of radar technicians who now had a bunch of time on their hands. There was also the opportunity to take advantage of a bunch of top-notch theoretical work done by Dutch astronomers during the war that led to the discovery of a radio emission line from hydrogen.
The basic design of a radio telescope is just like an optical reflector. One has a reflecting surface, and an instrument package at the focal point. But the telescopes are much larger. In part, this is due to the diffraction-limit issue, but it is also driven by the relative faintness of astronomical radio signals. As an example of this, the unit adopted to measure such signals is the Jansky. It is defined as
Power per unit area, per unit frequency. Note the exponent. Big dishes are essential to collect enough photons. The bigger the better. The apotheosis of this concept is the Arecibo radio telescope. Arecibo is constructed from the caldera of an extinct volcano in Puerto Rico, as has an aperture of 305m. This is, by far, the largest collecting area of any telescope in the world. Unfortunately, one cannot point a mountain. As a result, there is a band of the sky that passes over Puerto Rico that has been surveyed to a much greater depth in radio wavelengths than has the rest of the sky.
The largest fully-steerable radio telescopes are in the 100-m class. The Byrd telescope, at NRAO in West Virginia, is the current state of the art. At much smaller sizes, millimeter-wave telescopes are another area where substantial advances have been made in the last 20 years. As one is dealing with wavelengths that are smaller by one or two orders of magnitude, one can get the comparable angular resolution from a 20-30m dish. But the figure tolerance is higher (natch).
Recalling my example of the angular resolution of a 30m dish, working at 21cm, notice that this implies a need for a 60km dish to reach 1" resolution. This is clearly impossible. Because of this, radio astronomers were forced to abandon the brute-force method, and actually think about the problem. This led to the development of radio interferometry.
The basic idea here is that one observers the same target with two telescopes that are seperated by some distance D. Then, by combining the information from the two telescopes, one is able to construct an image of the source that has a resolution equal to what you would get with a single telescope of aperture D (although not the collecting area!). Thus, by building a network of telescopes (Image courtesy of NRAO/AUI), one can achieve baselines of 5000 km or so. Let us return to our earlier example of an observation at 21cm. Now with a baseline of 5000 km, we will get an angular resolution of
A substantial improvement!
Interferometry is really MUCH more complicated than the simple sketch provided above. Without getting too lost in the details, it is worth considering some of these complications. First, a simple 2-element interferometer will provide information on the angular scale determined by the seperation of the elements, but not very much information on either larger or smaller scales. Second, it will provide that information only at the orientation-angle provided by the two antennae. In practice, this means one wants more than two antennae, and one wants the pairs of antennae to each sample different scales at different orientations. One can take some advantage of the rotation of the Earth to extend scales and cover orientations that a "snapshot" would miss. Note that in an ideal case, the number of seperate baselines sampled by n antennae is
The atmosphere is opaque at x-ray wavelengths, so the field of x-ray astronomy is only just over 40 years old (the 1962 sounding-rocket discovery of Scorpius X-1).
A second issue for x-ray work is how to focus x-rays. One typically talks about this problem in terms of x-rays being "penetrating radiation", but it is more useful to think about the matter a bit differently. X-rays have wavelengths of around 1nm. But a typical solid has an interatomic spacing of about 0.1nm. In other words all solid surfaces are intrinsically rough (with inhomogeneities larger than 1/20 of a wavelength). Thus direct-incidence mirrors generally cannot provide good images in the x-ray. Beginning in the late 1970s, this problem was addressed by adopting grazing incidence mirror designs. In such a system, the x-rays make some small angle with the surface of the mirror. For a grazing angle of some theta, the resulting apparent roughness of the mirror will be reduced from x as follows:
Thus, for a small enough angle, one can obtain an apparent figure of less than 1/20 of a wavelength.
Up until the 1990s, the detector of choice for x-ray astronomy was the Proportional Counter, adopted from the particle physics community. The proportional counter has some intrinsic energy resolution, so one could do crude spectral energy distribution work. But this was a far cry from true spectroscopy. The ASCA satellite was the first x-ray mission to adopt the modern strategy. ASCA used CCDs as detectors.
In the optical, CCDs provide no inherent energy resolution, as the band-gap energy is roughly the same as the photon energy (ignoring issues of gain and such, one gets one photoelectron per detected photon, no matter if the photon is blue or red). However, as x-rays are so much more energetic than optical photons, the number of photoelectrons is a function of photon energy. For bright sources, it is also possible to obtain useful information from x-ray grating spectrometers. This has led to a revolution in x-ray astrophysics, as the quality of the observational data are now substantially better than the quality of the state-of-the-art atomic physics used to analyse the data.
The grazing incidence technique is inadequate for gamma rays, thus we still have no true gamma-ray imagers. The most recent space-based gamma-ray observatories are the Compton Gamma Ray Observatory (early 1990s) and the European Beppo-SAX mission from the late 1990s. The angular resolution of the gamma ray instruments on these satellites were quite poor by optical standards (10s of arcminutes at best).
At the very highest energies, one can use indirect ground-based telescopes to detect Cherenkov radiation from gamma rays striking atoms in the upper atmosphere. Such facilites have the decided advantage that they do not fall out of orbit and incinerate in the Earth's atmosphere.
The story of gamma-ray astronomy is the story of gamma-ray bursts. Gamma Ray Bursts were discovered in the 1960s as a by-product of a Cold-War era satellite program to monitor Soviet compliance with the atmospheric test ban treaty. The Vela satellites were equiped with gamma-ray detectors, to look for air-bursts from nuclear explosions.
Right off, they started detecting bursts every few days. But the detectors had no positional resolution. By the late 1960s, it was clear that the bursts were not coming from the ground, but were instead extraterrestrial.
The program was classified until 1973. Upon declassification, this is what was known:
First, it provided a sample of thousands of GRBs. This was important, as it demonstrated clearly that the bursts were isotropic. They showed NO sign of any concentration toward
There was another result from GRO that was significant to this discussion. That is that, although the distribution of bursts was isotropic, it was NOT homogeneous. This follows from counts of the number of bursts as a function of fluence. The calculation is outlined in the text. Essentially, for a homogenous distribution of sources that are all drawn from the same population, one expects
The observations clearly showed that there were many fewer faint bursts than expected, indicating that the population had an edge. This ruled out the remaining Galactic Plane models right off.
The story has changed enormously about ten years ago due to observations with the Dutch/Italian x-ray/gamma-ray satellite Beppo/SAX. Beppo/SAX reported an x-ray counterpart to GRB 960720 with a positional accuracy of a few arcseconds.
Realizing this was possible, the observatory team concentrated their efforts on quickly finding, and localizing GRBs. The result of this is that x-ray, optical, and radio counterparts of GRBs were soon found (GRB "afterglows").
The current state of affairs is that GRBs are known to be associated with type II SNe events in spiral galaxies. The optical spectra of GRB afterglows are consistent with SNe IIs, and the events are seen in spirals, associated with regions of recent star formation.
One can detect cosmic rays both by traditional particle-physics means (stacks of emulsions, proportional counters, and such), or by using Cherenkov radiation telescopes. Cosmic rays are particles of matter (mainly protons, but also heavier nuclei) that are excited to relativistic velocities by high-energy events. The standard understanding is that supernovae (exploding stars) are responsible for most of the cosmic-ray production. Such models are fairly good at matching the observed distribution of cosmic-ray energies except at the very high-energy end. At these highest energies, the models predict the flux of cosmic rays should drop to zero, but the observations indicate a flattening (or even a rising) in the energy spectrum. The problem with this is that the cosmic rays are trapped by the magnetic field of the Galaxy. The drop-off should occur at the energy that allows cosmic-rays to escape from the Galaxy. Why the spectrum appears to flatten out at that energy range is an outstanding mystery.
There are two basic classes of neutrino observatories:
To date, there have been two astronomical objects with detected neutrinos: The Sun, and Supernova 1987A. As you may know, the Solar neutrino problem was a 35-year mystery. The basic problem is that, beginning in the mid-1960s, the measured Solar neutrino flux was about 1/3 of the predicted flux based on Solar modelling. The resolution of this problem occured only a few years ago, and has had major impact on our understanding of particle physics. The issue turned out to be that neutrinos can change flavor. The Solar neutrinos that earlier experiments measured were electron neutrinos. The "missing" ones had transformed into either muon or tau neutrinos on their way out of the Sun. The impact on particle physics comes from the realization that such flavor mixing can only happen if neutrinos have a non-zero rest mass.
Supernova 1987A occured in a nearby galaxy. At that time there were several operational neutrino detectors, and two of them saw excess neutrinos at the time of the supernova explosion. The total number of detected neutrinos was on the order of 10. The total number of papers published on this event is well over 100.
Hopefully, the next time a supernova goes off within 1 Mpc or so, the above ratio will be a bit less drastic.
We have now arrived at the limit where the ratio of publications to detected sources is infinite. To date, there have been no direct detections of gravitational radiation. There has been one indirect detection: The binary pulsar system PS 1913+16. A neutron star-neutron star binary system is a wonderful laboratory for Gravity physics. This is because one has two very dense mass concentrations in short-period orbit around one another. This causes a periodic distorsion in the surrounding spacetime that results in gravitational radiation. Such radiation is a means of carrying energy away from the system, and will cause the members of the binary to spiral in toward one-another. The system PS 1913+16 is just such a binary system, and has been the subject of intense scrutiny for the past 25 years. The properties of this system offer the best confirmation of General Relativity in the strong gravity limit that have been observed to date.
The search for gravity waves has used a number of approaches over the years. The one that offers the current best-bet is LIGO (Laser Interferometer Gravity Observatory). LIGO is now just coming on-line. It is a pair of two-arm laser interferometers, seperated by ~2000 km (one is in Hanford WA, the other near Baton Rouge LA). The simplest sketch of the idea is that each of the two perpendicular arms of a given interferometer will be most sensitive to gravity waves that compress and stretch spacetime along a given arm. Thus one looks for differences in the interference patterns along the two arms, and uses the distant pairs of interferometers to cull out local phenomena (trucks, earthquakes, etc.). If LIGO does NOT succeed in detecting gravity waves, then our ideas about gravity physics are going to require some major revisions
