What are you looking for? Topography? –Is the surface dirty? –Are you looking for the clean surface or the dirt? –If the dirt, no preparation is required! –If the clean surface, dirt must be removed Solvent in ultrasonic cleaner Methanol is best to finish cleaning Plasma cleaning also an option

Ultrasonic cleaner

Distribution of Chemical Elements EDS/WDS analysis –“Flat, smooth and polished” –Ideally smooth on the micron scale Backscatter determination of chemical inhomogeneities –Less demanding of polishing –Failure to polish can lead to artifacts

Sample Preparation Scanning Electron Microscope Beam of electrons rastered across sample Current flows from column to sample How does it complete the path back?

There are two kinds of microscopes: 1.High vacuum 1.“Requires” samples to be conductive 2.Alternative: reduce beam energy to the point where current of secondary and backscattered electrons equals probe current 2.“Variable pressure” 1.Backfill gas bleeds off charge 2.Backfill gas amplifies signal 3.Allows imaging of non-conductive materials

Start simple 1. Conductive samples: metals, semiconductors… –Fasten sample directly to stub –Carbon tape: 8 mm to 50 mm wide x 20 m –Carbon dots: 9 mm, 12 mm, 25 mm –Carbon sheet: 5 x 12 cm –Aluminum/Copper tape: single or double sided –Silver paint –Carbon paint

Carbon dot: 200 X

Carbon dot: 10,000 X

Coating of non-conductive samples Sputtering of gold film –Surface typically pre-etched with plasma –Gold coating is nominally 10 nm thick –Sputtering should give conformal coating –Au/Pd preferred for smaller particles Carbon coating –“High” vacuum deposited: coating not conformal: works best on polished samples –Simple thermal evaporation of carbon from yarn –Preferred for analysis for light elements

Low Z material - polymer, or biological Low BSE coefficient Large exitation volume High BSE coefficient Small exitation volume BSE PE R SE II The wrong way to coat  Coatings can be THICK or THIN, particulate or smooth  Coating are usually high Z metals such as Cr, Ta, W, Pt, Au  With a THICK (20-50nm) coating the beam interaction occurs mainly within the coating Layer  The SE-signal is then mostly SE2 (i.e. converted BS electrons)  The topographic resolution is limited by the thickness of the metal coat and the SE II range (i.e ~  m) This slide stolen from David Joy

SE II SE I BSE R LLBSE The Right way to coat  Use a THIN film  The beam interaction is now mainly in the sample  The SE-signal is SE I from the coating  There is very little SE II from the metal layer.. and no signal from specimen itself because SE produced beneath the metal layer cannot leave the specimen  Topographic resolution is now only limited by thickness of the metal coat and the diameter of the electron beam

Particulate Coatings  Au produces very big particles (30nm) so do not use pure gold for coatings  Au/Pd, Pt, W, or Ir make smaller (<1-3nm) particles  All have a very high SE yield and can be deposited in a low cost, sputter coater  The coatings are stable for long periods of time  Particulate coatings are ideal below 100kx but they can be useful even at higher magnifications.. 3nm layer of Au/Pd

UHR SEM Coating Results Uncoated Pt coated Hitachi S5200 Note the benefits of a reduction in charging and the gain in image contrast and detail. The fine grain - while visible - permits accurate focus and image stigmation. Resolution ~ 1.3nm Courtesy Bryan Tracy (Spansion)

How to make good particulate coatings  Whether they are to be used for high resolution or charge control the same general rules apply to all metal coatings  The basic items of equipment required are:  A sputter coater, or a ion beam unit  A clean gas supply, safely fixed to the laboratory wall and carefully plumbed to the gas inlet on the coater  some clean pieces of filter paper  patience

Four good rules to follow  Keep it clean - wipe the glass vessel clean after every run, and clean the anodes weekly. To keep the glass clean wipe it with a drop of liquid detergent before use and then clean with a lint free cloth.  Keep it slow - reduce the gas pressure and/or the anode voltage till the plasma just stays on  Keep it dry - the argon gas (never air!) must be perfectly dry. Check that the color of the plasma is deep purple (indigo) rather than pink  Keep it thin - as we saw earlier thick coatings lower the resolution of the image by generating more BSE and SE2 electrons. The actual thickness needed depends on the material being used and cannot be checked reliably by film thickness monitors

Quality Assurance  Particulate metal coatings  Aim for a single layer coverage - thickness will then be equal to the average grain size, typically 1.5 to 3nm depending on the metal used  Checking - the shadow on the filter paper should only be visible under strong, oblique, illumination and will be a ‘blush’ of gray. Any hint of color, or a metallic sheen, indicates that the film is nm thick at least

Mass thickness contrast If the metal layer is smooth and continuous then a new and important benefit occurs for high resolution imaging The SE1 varies with the thickness of the metal film This effect saturates at a thickness equal to about 3λ The conformation of the film to surface topography thus provides contrast 1nm2nm3nm Film thickness SEYieldSEYield bulk value mass thickness variation

Metal builds contrast  All of the SE1 signal comes from the metal layer  The resolution will be of the order of the layer thickness  The mass thickness effect gives extra contrast enhancement at the edges because of the increase is projected thickness  The feature is now truly ‘resolved’ since its size and shape are visible once more 5nm low Z object 2nm metal film Beam position SE profile with metal film SE profile without metal SESE

50 nm Imaging Low-Z materials with mass contrast coatings From René Hermann et al., (1991) J Struct Biol 107: FabFragments  In this example a 1.5nm Cr film has been used  The mass thickness contrast resolves edges and make nanometer scale detail (such as the molecular Fab fragments) readily visible  The high SE yield also improves the signal to noise ratio of the image  Unfortunately - use at once then discard because Cr films oxidize and degrade very rapidly in air

Depositing Cr films  Useful Cr coats cannot be deposited in a rough-pumped sputter coater because of the rapid oxidation that occurs  The system should have a base pressure of T and be dry pumped. After initial pump-down back fill the system with dry nitrogen to a pressure of about 1 Torr. Pump down once more to the base pressure, then back fill and pump again. Finally fill the liquid nitrogen trap to lower any residual water vapor pressure to below T  For the first 60 seconds of deposition the sample should be covered by a shutter to prevent oxide reaching the surface  By experiment determine the slowest possible deposition rate. This ensures the smoothest and most uniform film

Structureless Coats  Structureless coatings  Aim for 1 to 1.5nm to get the maximum SE1 contrast across features. These films are hard to focus and stigmate and intensity is low  Checking the film thickness is impossible. Try varying the time until the best image is produced then copy exactly  Remember that Cr, and Ti are very reactive and will oxidize rapidly even in a good vacuum. Once deposited these films should be used immediately and then discarded. Other materials may be more stable

Coating Summary  Coatings are an essential part of the technique of high resolution SEM because they generate interpretable contrast, improve resolution, and enhance the S/N ratio  Thin coatings are better than thick coatings - do not make your sample a piece of jewelry  Below 100kx magnification particulate coatings are superior to Cr or Ti because of their higher SE yields  Above 100kx magnification can use chromium or titanium continuous films to generate mass thickness contrast and enhance resolution, or can use nano-grain Pt or W films (even though the grains are visible) for charge control and as an aid to accurate focus and stigmation  Carbon is a contaminant not a coating unless it deposited by an ion sputter tool

Plastic casting: the recipe Mold: Heavy wall aluminum tubing, 1.25” ID (matches fixture on polisher) Coat inside with Buehler release agent Cover end of tube with large mailing label! Add/orient big pieces with tweezers; dump in powder Epofix resin and EpoAR hardener from Struers

Plastic casting: the recipe cont’d Mix epoxy and hardener in disposable beaker Stir sample and powder with toothpick! –Mix very thoroughly –Evacuate Watch foam rise When in danger of overflowing, backfill Repeat three times to get rid of bubbles Works well in a plastic dessicator so you can see what you’re doing

Plastic casting: the recipe cont’d Leave at 40 °C in open air overnight Drive plastic cast out of tube with mandrel Flatten surface to be polished with belt sander Polish with successively finer grit pads –45 micron diamond –15 micron diamond –6 micron diamond solution from Struers –3 micronon Buehler pad –1 micron

Plastic casting: the recipe cont’d At each stage of the process –Clean the casting with water in an ultrasonic cleaner to avoid carrying coarser grit to finer pad –Observe the surface with an optical microscope –Have scratches from previous (larger) grit been polished out?

Plastic casting: the recipe cont’d Finished product is flat, smooth and polished on scale of 1 micron –Scale is comparable to sampling volume for electron-induced X-rays –Absolutely mandatory for high accuracy WDX measurements of chemical composition Technician in Geology whose job is to do precisely this sample preparation process