FIKl.D E\IISSIO> MICHOSCOPY 



chanies. Electrons inside a metal are moving tails to variation in work function. Some- 



iii a potential trough, from which they can times, however, it is evident that small crys- 



escape by adding thermal energy (thermi- tallites or asperities appear in the image 



onic emission), by energy transfer from pho- because of the local field enhancement. 



Field Emission Patterns from Clean 

 Surfaces 



tons (photoelectric effect) or by penetrating 

 the potential barrier at the surface (tunnel 

 effect) when the external field is so high that 



the barrier width is narrowed down to be There is hardly another experiment that 



comparable with the electron wavelength shows as clearly as field electron microscopy 



inside the metal. According to the Fowler- the difficulty of obtaining and investigating 



Nordheim theory of field emission the current really clean metal surfaces. The number of 



density J (amps/cm-) as a function of field gas molecules striking a square centimeter of 



strength F (volts/cm) and work function surface in a conventional high vacuum of 



(e-volts) at the surface can be described by 10~^ mm amounts to 4 X 10^^ per sec, and 



P2 as the sticking probability on a clean metal 



/ = 1.55 X 10-6 — e-(« 86X10 *' )/F, (1) gui-face is near unity, it takes only a few 



seconds to build up a monolayer of adsorp- 



If in a more elaborate theory the electronic tion at that pressure. Therefore, studies of 



unage force is taken into account, the expo- clean surfaces or of adsorption films under 



nent will be reduced by a factor slightly reproducible conditions require ultra high 



smaller than unity dependent upon ■\/F/<t>- vacuum techniques. In field electron micros- 



This reduces the field strength necessary for copy the criterion for a clean surface is that 



a given current density by some 10 to 20%. the emission pattern should be independent 



As field emission is essentially independent of the annealing temperature except for small 



of temperature, the behavior of the emitter changes in the relative size of the dark 



surface can be observed in a wide tempera- areas. 



ture range from liquid helium temperature Only a few crystallographic planes of low 



to white heat, which is quite valuable for the Miller index appear on a clean surface as 



study of adsorption. dark regions, indicating a higher work func- 



The resolution of the field emission micro- tion than the rest of the surface. Probe meas- 



scope is limited by the random lateral veloc- urements of the current density in various 



ity component of the emitted electrons, and regions revealed large differences in work 



also by diffraction of the electron waves. At function. On tungsten 4>\n was found to be 



a tip radius of 1000 A the theoretical resolu- 4.31 e-volt, and (^on as high as 5.99 e-volts. 



tion is 20 A, and at 3000 A radius, 40 A. This Observ^ations of field electron microscope 



is in agreement with practical experience. At patterns of other metals strongly suggest the 



small protrusions at the tip surface the reso- general occurrence of such previously not ex- 



lution may be better due to local field dis- pected large work function differences be- 



tortions. The image details or the contrast tween different crystal planes. On body-cen- 



are the result of local variations in current tered cubic crystals, such as W, Mo, Ta, 



density, which according to Eq. (1) is deter- aFe, the highest work functions are on Oil, 



mined by the local field F and the work func- followed by 112 and 100 planes. Face-cen- 



tion <f). The interpretation of the image is tered cubic crystals, such as Ir, Pt, Ni, have 



usually based on the assumption of a homo- the highest work function in 111 and 100 



geneous field strength over the depicted tip planes, and on hexagonal Re crystals the 



area, falling off toward the shank of the highest work function is on the 0001 plane, 



emitter, and by ascribing the finer image de- followed by 1010. 



326 



