PRODUCTION OF CONTINUOUS AND CHARACTERISTIC X-RADIATION 



anisotropies for thick and thin targets. It ap- mann's data. The agreement is in this case 

 pears in fact that multiple scattering and dif- within the experimental error, 

 fusion prevail even when the electrons have More recently the theory has been sub- 

 lost only a small fraction (one tenth in this jected to a relatively comprehensive test by 

 example) of their initial energy.- This implies the work of Amrehn and Kuhlenkampff 

 that virtually all the x-ray output from a (1955) and Amrehn (1956). The latter au- 

 thick almninum target for an incident elec- thor has made a detailed absolute com- 

 tron energy of 10 keV is produced from elec- parison between his experimental energy 

 trons Avhich are fully diffused. "Full diffu- distributions for five elements and the calcu- 

 sion" and its relation to multiple scattering lationsof Kirkpatrickand Wiedmann. Agree- 

 is discussed by Bothe (1932) and Paul and ment is often better than 10% and nearly 

 Steinwedel (1955). always better than 20%. The general im- 

 pression is that the theory is now well veri- 

 Efficiency of Production of Continuous ^^^ ^^^ bombardment energies in the region 

 X -radiation ^f 20-40 KeV. 



The main source of theoretical information Kirkpatrick and Wiedmann extended their 



regarding the efficiency of production in calculations to give an expression for the 



thin targets is the data of Kirkpatrick and thin target efficiency of production of x-radi- 



Wiedmann (1945), and for comparison some ation, by integrating their energy distribu- 



experimental data are available. tion curves over all energies and directions, 



Massey and Burhop (1952) have con- and by comparing this with the energy loss 



sidered the data of Smick and Kirkpatrick suffered by the bombarding electrons. The 



(1941) and Clark and Kelly (1941), con- thin target efficiency is given approximately 



verting it into absolute units where necessary, by 



and drawing upon Kirkpatrick and Wied- = 2 8 X IQ-^ZV (V in volts) 

 mann's data in preference to the earlier cal- 

 culations used by the experimenters. Smick although the calculations show that it is not 

 and Kirkpatrick measured the radiation from strictly a function of ZV, but depends upon 

 a thin nickel target bombarded by 15 keV these two variables separately to some small 

 electrons, using balanced filters to select a extent. 



band of radiation of quantmn energy about Turning now to thick targets, the experi- 

 8 keV or 0.575 of the high energy limit. At "^^ntal data were reviewed by Compton and 

 88 degrees to the forward direction the in- ^^"^^^^^ ^1935), and the expression 

 tensity was 2.2 X 10"^" ergs/steradian/unit ,, = 1.1 X lO'^ZV (Fin volts) 

 frequency interval/electron/atom per cm^ ^^^^^ eonsidered to be accurate to about 20 %. 

 The data of Kirkpatrick and Wiedmann yield ^^.^ eonfirmed the calculation of Ivramers 

 a value of 6 X 10-^« in the above units, al- ^^^ obtained the result 

 though Smick and Kirkpatrick give a calcu- 

 lated value (after Sauter (1934)) of 2.9 X ' v = 0.92 x KT^ZV 

 10-^°. Clark and Kelly made a similar meas- ^^.^^ j^g theoretical expression for the energy 

 urement using alimiinum at a bombarding distribution from thick targets. Kirkpatrick 

 energy of 31.7 kV and a quantum energy of and Wiedmann calculated the efficiency by 

 0.82 of the high energy limit; at 60 degrees integrating their thin target data over all 

 they found the intensity to be 6.2 X 10~^^ angles and energies and found the efficiency 

 ergs, etc., (±33%), as compared with to be proportional to voltage and atomic 

 4.23 X 10- *i from Kirkpatrick and Wied- number to quite a high degree of accuracy, 



659 



