Surface Emissions Pole Piece, etc SE3 X-rays Cathodoluminescence ≈ 1 nm for metals up to 10 nm for insulator Specimen current Dependence of backscatter coefficient on backscattered electron energy Backscattering increases strongly with Z independent of beam energy Backscatter dependence on tilt and Z Where are electrons coming from? • Kanaya – Okayama equation for electron range in a material: • Rk-o = (0.0276 A E01.67)/(Z0.89ρ) μm – For E0 in keV, ρ in gm/cm3, A in gm/mole and Z = atomic number Where are electrons coming from? Material A Z ρ E0 R (μm) PMMA 12.01 6 1.16 20 8.63295 Fe 55.26 26 7.87 20 1.58759 Fe 55.26 26 7.87 1 0.01066 Homework 2.2 Depth of origination of backscattered electrons Backscattered electrons can Investigate deep into the sample Distribution of scattered electron energies Note that the distinction between SE and BSE is a definition There are only “scattered electrons” Why do backscattered electrons give the most information about chemical composition ? Why don’t secondary electrons give much information about chemical composition? Dependence of scattered electron yield on Z 50%! Nearly independent 10% Why do secondary electrons give great topographic information? Image formation and interpretation Image formation and interpretation Images from exactly the same area of the same sample taken with different detectors. In case you thought the second image was just taken at higher contrast… And it’s not just your detector choice that can impact your image… 20 kV 500 V Everhart- Thornley detector (1960) Scanning and Data Collection Transfer of image from sample to screen Works for both topographic and elemental information Pixels (Picture elements) • NEC monitor: 380 x 300 mm; 1280 x 1024 pixels. Hence pixel size on monitor size is 297 x 293 microns...300 microns. • Typical file size used in FEI Nova NanoSEM is 1024 x 884 pixels. • Pixel size on sample is pixel size on monitor divided by magnification, about 15 microns (20X) to 0.6 nm (500kX). So what; who cares? • Example: 10 keV beam at 100 pA viewing at 100 X (neural array was taken at 118 X) saved into 1024 x 884 file using Leo and FEI. • β = 4Ip/π2αp2 dp2 • dp = (2WD/πrA)sqrt(Ip/ β) • FEI: dp = (2*5 mm/π*.015 mm)sqrt(10-10 Acm2sr/108A) = 0.5 nm! • Leo: (2*8 mm/π*.01 mm)sqrt(10-10 A cm2sr/105A) = 40 nm! So what; who cares? Cont’d • So – Your pixel size varies from 0.6 nm to 15 microns – Your beam diameter can vary from 0.5 nm to 40 microns, at the smallest – Your interaction volume varies from 10 microns to 10 nm (BSE, SE2; last homework) Implications • If probe size is too small – You are wasting resolution: topography can change between sampling points (Nyquist Theorem!) – Resolution regained by sampling and saving more points – You are wasting signal to noise – You are wasting contrast How to make the spot larger? • dp = (2WD/πrA)sqrt(Ip/ β) – Increase working distance – Go to a smaller aperture – Increase probe current – Decrease accelerating potential • Is this the dominant effect of decreasing the accelerating potential?
Pages to are hidden for
"Image formation and interpretation"Please download to view full document