296 ANNTJAL REPORT SMITHSONIAN INSTITUTION, 1963 



seeing may approach 0.2 to 0.3 second of arc, as has been noted at 

 Pic du Midi in France ; however, this size is still much larger than the 

 theoretical resolving power of a large instrmnent. As a consequence 

 a very large telescope can promise only a larger picture of the same 

 blurred celestial object as would be obtained with one perhaps only 

 one-half as big. 



The second possibility for improvement is in the efficiency of the 

 detection of the photons. The photographic process, widely used for 

 many years, has the ability to record stars over a wide range of 

 brightness, although the accuracy of the measure of brightness is 

 relatively low. One photograph may record star images over a 

 range of 20 magnitudes (10^ in intensity) and also record a million 

 information elements per square millimeter. Information densities 

 up to 5 million elements per square millimeter are possible under 

 laboratorj'- conditions. The quantum efficiency of a photographic 

 emulsion is low, ranging from 0.1 to 1.0 percent. There is little hope 

 for a large improvement in the photographic process itself since 

 individual silver grains in the emulsion are quite good detectors. The 

 quantum efficiency for a single gi'ain to be developable in terms of 

 absorbed photons is 25 percent. One developed grain, however, does 

 not provide a detectable quantity since every grain produced by 

 chemical reaction called "fog" would be indistinguishable from a 

 "star". Only when groups of 20 or more grains are developed does 

 one recognize the clump as an entity on the background of fog grain 

 dumpiness. 



In recent years much effort has been devoted to the utilization of the 

 high quantum efficiencies approaching 30 percent for the photoelectric 

 detector. The photomultiplier is a commercially available device of 

 high efficiency and built-in amplification which has been widely used 

 in astronomy and nuclear physics. The internal amplification of 

 such a device of 10^ produces a measurable pulse each time a photo- 

 electron is emitted from the cathode. The cathode will occasionally 

 reject a "thermal" electron spontaneously as a consequence of the 

 low work function of the caesium compound emitting surface. These 

 thermal electrons produce what is called the "dark current," which 

 adds a noise background to the signal. A good photomultiplier at 

 room temperature will have a dark current of 10 to 20 electrons per 

 second from a 1 cm.^ photocathode surface. Because the dark current 

 emission is temperature dependent, the astronomical use of photo- 

 multipliers for use on faint objects is always with the device cooled 

 to dry-ice temperature ( — 80°C). At this low temperature a good 

 tube will have a dark current of about 0.2 electron per second per cm.^ 



The photomultiplier is an excellent detector of a single object at a 

 time. The output current is accurately linear over a wide range of 

 intensities; hence, the brightness of a star can be measured veiy 



