ULTRASOFT X-RAY MICROSCOPY 



A, ALUMINUM K(,-8.34 A 



B. OXYGEN Ka-23.7 A 



J I I i_ 



I I I I I I I I I I I I I I 



NOISE- N 



-f^^^ 



■•* H 



I I I I I I i I I I I I I 



120 



80 



40 





 12% 



8 



4 







3 



y/N 



J I I I L. 



3 



LOG E- 



DENSITY-D- 



FiG. 14. The D vs log E characteristic for two wavelengths illustrating the longer linear region at 

 the higher densities for the ultrasoft wavelength and the associated concontrast values, 7, and relative 

 granularity noise, dp/p, measured with 0.5 X 1.0 micro slit microphotometry. 



rapidly with wavelength than does (y/N), 

 indicating the value of the longer wave- 

 lengths in reducing measurement error. For 

 a given wavelength, the value of {y/N) may 

 be considered a "figure of merit" for micro- 

 radiographic emulsions. The value of (y/N) 

 is a maximum in the linear high density re- 

 gion is thus the optimum working region for 

 both low and high magnification microradio- 

 graphic measurement. 



(y/N) has been measured in a recent in- 

 vestigation (9) as a function of the type and 

 time of photographic development which has 

 indicated that a "soft" development such as 

 2 to 3 minutes in D158 (Eastman) seems 

 optimum. 



We may now summarize, from the fore- 

 going discussion, the procedures for an opti- 

 mum photographic measurement in mi- 

 croradiographic analysis: The absorption 

 parameter, /xm, may be simply determined 

 for either high or low magnification through 

 the relation 



fxm = (I/7) In (pi/pi) 

 providing the exposure is for the linear high- 



density region (D > 1.0 for ultrasoft radia- 

 tions and spectrographic-649 emulsion) for 

 which 7 is a constant and of maximimi value. 

 The minimiun wavelength permissible for 

 low magnification (dpi ;^ dp2) is set by the 

 requirement that yiiin should lie within the 

 range 0.7 to 2.0. The measurement errors 

 become very large in this case for yum < 0.5. 

 The minimimi wavelength is set for high 

 magnification (dpi/pi ^ dp2/p2) by the re- 

 quirement that ij.m satisfies the relation (15) 

 for a given required measurement precision, 

 d(fj.m)/fjLm, (or for dm/w-uniform chemistry) 

 and for a given emulsion and hence (y/N) 

 value. The minimum wavelength and ex- 

 posure time are thus fixed for each microra- 

 diographic measurement. 



As shown, for example, in Fig. 14, the 

 (y/N) value for Spectroscopic-649 plate for 

 23.7 A radiation is about 60. Therefore, from 

 (15), for a minimum error 2%, the x-ray 

 wavelength should be chosen so that juw is at 

 least equal to unity, corresponding to ap- 

 proximately 60 % absorption. 



To measure high densities on microradio- 

 gram areas of about one square micron or 



689 



