o76 FRED M. U B E R 



ammeter A3. The electrons emitted by tlie filament traverse box N, as indi- 

 cated by the dotted line, owing to an accelerating potential of 90 v. supplied 

 bj^ battery Bt. Most of the electrons are trapped on plate Q, \^•hich is about 

 90 V. more positive than N. The electron current to Q, perhaps 20 ^uamp., 

 is read on meter A5 and the total electron current to both Q and A'', perhaps 

 40 iuamp-, is read on A 4. A magnetic field in the direction of the electronic 

 flow, provided by the magnet M2, either permanent or electric, confines the 

 electrons to a rather well defined beam. 



The ionized particles produced by the electronic bombardment acquire 

 a small velocity in the direction of slit Si from the accelerating voltage sup- 

 plied by ])attery Bt. But the principal accelerating potential difference of 

 nearly 1000 v. is furnished by battery Bi or a suitable rectifier (I4), and oc- 

 curs between exit slit Si in box A^ and entrance slit »S2 of C. In traversing 

 the length of the long copper tube, C, the ions are deflected 60° by the sec- 

 torial magnetic field of the electromagnet, Mi, and brought to a focus at 

 slit S3. Upon emerging from C, the ions are collected on plate P. To pre- 

 vent scattered electrons from the edges of slit S3 from reaching P, a small 

 negative potential from battery B^ is applied to an intervening slit plate. 



Because of the charged particles collected by P, and the resultant current 

 through R, the grid of electrometer tube E experiences a change in potential. 

 The resistor, R, has a resistance in the neighborhood of 10^" ohms. Usually 

 two or more such resistors having different values are arranged so tliat either 

 may be used at will in order to change the sensitivity. The calibrated 

 variable amplifier, //, may include a circuit as described by Penick (9) or 

 perhaps an inverse feedback type of amplifier {4,6,13). A wide range of 

 sensitivities is desirable since relative isotope concentrations may vary by a 

 factor of several hundred. In addition an automatic switching arrangement 

 (5) should be provided to facilitate altei'uating the observations rapidly be- 

 tween the two or three mass peaks that characterize a particular test gas. 



The operating gas pressure is less than 10 ~^ mm. of mercury, and should 

 be read by an indicating gage, / {10,12), or by a McLeod gage located be- 

 tween trap Ti and diffusion pump Di. The trap is kept cold with either a 

 solid carbon dioxide mixture or liquid air, depending on the material being 

 tested. Mercury vapor diffusion pumps seem to be most satisfactory. Small 

 rotary oil vacuum pumps, Oi and O2, precede the diffusion pumps. 



Magnet Mi may be wound for operation from either 12 v. storage batter- 

 ies or a rectifier (4), or it maj^ be replaced bj^ a permanent magnet {6,8). 

 Electronically regulated power supplies may be substituted for batteries in 

 any of the indicated circuits. To facilitate baking out the spectrometer tube, 

 the main magnet. Mi, is mounted on a carriage. It may be rolled to the left 

 in the figure, thus enabling heating elements to be placed in position around 

 metal tube C. Kovar collars, K, facilitate copper to Pyrex Vacuum seals. 



Most mass spectrometers built have had deflection angles of 



