use of field emission electron microscope 925 



Formation and Disappearance of Crystallites 



As the number of Ba shots increases in Fig. 7, a large dark region develops 

 and enlarges from the right 100 region. This is due to the well known fact 

 that when the Ba concentration exceeds one monolayer the work function 

 increases. If the Ba atoms remained where they were deposited, one would 

 expect the patterns to consist of broad bright arcs whose centers were at the 

 region of greatest deposition and whose radii would increase with amount 

 deposited ; for 7 shots of Ba one would expect a narrow bright arc of nearly 

 maximum possible radius. Such an arc does indeed appear after 6 and 7 

 shots; but the regions with more than a monolayer are not dark as expected; 

 instead there appear in these regions intensely bright large area emission 

 centers. These just begin to show after 3 shots and become more prominent 

 for 4 to 7 shots. These bright emission centers have properties different 

 from those of the clusters previously described; they are larger, may ap- 

 pear on any plane, are in a continuous state of flicker even at 300°K, dis- 

 appear at much lower glowing temperatures and can be observed at much 

 lower applied voltages. In accord with Haefer,^ we believe that they are due 

 to Ba crystals which grow normal to the surface; hence the term crystallites 

 seems appropriate. At the crystallites the local field should be much greater 

 than the average field and hence they should be observable at low applied 

 voltages. Different crystallites should have a range of sizes and hence a range 

 of spot sizes. Crystallites should occur only for Ba concentrations greater 

 than monolayers and hence should be nearly independent of the underlying 

 tungsten. Because of the very high current densities through a crystallite, 

 one would expect a considerable increase in local temperature, perhaps 

 even to the melting point of Ba which would change the size and shape of 

 the crystalHte and hence the emission; this accounts for the flickering and 

 agrees with the observation that the amount of flickering increases with the 

 applied voltage and emission current. If the crystalHtes are solid Ba they 

 should evaporate more easily than Ba clusters adsorbed on W. Furthermore 

 the existence of similar crystallites for evaporated films has been deduced 

 from electron diffraction experiments. Hence the evidence for crystaUites is 

 quite good. 



Figure 8 shows the evidence for the disappearance of crystallites in the 

 temperature range 370 to 615°K. Note that, as T increases from 370 to 

 510°K, the voltage required to get 50 microamps decreases from 3150 to 

 2500. Note also that up to 615°K no detectable amount of Ba migrates into 

 the region beyond the outermost arc. The intensity of this arc decreases, 

 presumably because V is decreased. 



