Scanning Transmission Electron Microscopy Facility
STEM Specimen Preparation
Assuming your specimen in solution is perfectly mono-disperse and stable, the quality of STEM data still depends on our ability to mount structures on the thin carbon substrate with appropriate spatial distribution and structural preservation. Therefore the STEM group has worked to develop methods, materials and reagents that take this from “black magic” to a reproducible science. Beth Lin maintains the highest standards in all steps of preparation. However, there are still uncontrolled variables that require a dilution series of at least 3 grids for each sample (to obtain optimum spatial concentration) and careful searching over multiple areas of each grid to locate the best and worst areas (to assess any problems which may be less obvious in apparently good areas).
The practical requirements for accurate mass mapping are:
The STEM group provides 2nm thickness carbon film substrates on holey film coated 2.3mm diameter titanium grids mounted in a 0.3mm thick titanium holder. The holder keeps the grids flat and permits handling with tweezers without damaging the grid. Titanium was chosen because it is strong and forms an extremely stable oxide coating on the surface that is chemically inert. This contrasts with copper, normally used for EM, which is soft and quite soluble in buffers, especially ammonium salts. Copper binds to nucleic acid and carboxylic acids, giving erroneously high mass.
The carbon film is prepared by vacuum evaporation onto freshly cleaved single crystal rock salt using a calibrated carbon arc in an ion-pumped (oil-free) bell jar system. Mica substrates can be used as an alternative, but the UHV vacuum of our bell jar removes water and organic contaminants from the mica surface, making the carbon very difficult to remove. Holey film is prepared by the Fucami-Adachi method, floated on water, picked up on grids, coated with thick carbon and treated with ethyl acetate to remove any plastic. The holey-film grids are then mounted in holders.
Thin carbon films have the annoying property that they bind air pollutants, becoming hydrophobic and useless for mounting specimens. Some investigators use glow-discharge treatment to render films hydrophilic, but we find that this leads to very tight binding of the specimen (and junk), frequently causing flattening and disruption. Instead, we avoid any air exposure of the binding surface using the “wet-film” protocol we developed.
In the “wet-film” method, the thin carbon is floated off its rock salt substrate (water soluble) by inserting the salt into the water at a shallow angle, then discarding the salt block. The thin carbon film is barely visible (with suitable illumination) floating on the water surface as a raft 1x2cm. Holey-film grids in holders are placed face down on the raft and allowed to adhere for several minutes. Then they are teased apart gently with tweezers.
The wet-film, hanging-drop technique requires clamping one floating grid in a tweezers by the edge, lifting vertically to retain a drop of water and placing the assembly on a stand with the grid still face down (and the water drop hanging downward). Approximately 3/4 of the drop is removed by wicking from the bottom edge with filter paper and replaced by injection from below with 2ul of fresh water. This withdrawal & replacement is repeated 2x with water, then TMV solution (allowed to attach for 1 min), injection buffer for the specimen (3x), the specimen itself (allowed to attach for 1 min) and 10 changes of volatile buffer (usually 20mM ammonium acetate, pH 7).
(If the sample contains detergent, the injection and washing are complicated by the tendency to flow through the grid to the back, as well as between the prongs of the tweezers. This compromises the washing procedure leading to some non-volatile residue on the grid and breakage or removal of the carbon substrate. Also, the presence of objects on both sides of the substrate complicates the analysis. If detergent is absolutely necessary, the number of grids per specimen should be at least doubled.)
Next the grid is quickly pinched between layers of filter paper and plunged into super-cooled liquid nitrogen (super-cooling prevents boiling and gives instantaneous freezing of the ~ 1u thickness of liquid remaining on the grid). At that point the grid can be maintained under liquid nitrogen indefinitely (in our “grid fridge”) until the freeze dryer is available.
As an alternative for negative staining, the final washing is performed with either 2% uranyl acetate pH 3 or 2% methylamine vanadate (nanovan) pH 7 and the grid is air dried after blotting. Grids with negative stain are unaffected by passage through a freeze-dry cycle with their companion unstained grids.
All these operations are very sensitive to contaminants. A single particle of dust falling onto the water surface while the grids are floating can compromise an entire batch of grids. Therefore all operations are performed in a HEPA-filtered laminar flow hood. Water quality is also critical. We routinely perform water controls by allowing and entire 5ul drop of water to air dry on a grid. To obtain residue-free water, we start with distilled water from the Biology Bldg. and pass it through a Millipore Milli Q system to remove ions (18MΩ conductivity). Particulates are removed by distillation through a Corning glass still operated continuously at 25% of rated power. Water collected during the day is discarded and water collected overnight is stored in 2-liter polyethylene containers. Ammonium acetate volatile buffer is made fresh weekly and adjusted to pH7 using high purity reagents.
Freeze drying takes place overnight in a separate ion-pumped system. A specimen exchange cartridge with slots for 7 grids is mounted horizontally on this system. A “trap door” seals the bottom of the rack of specimen holders, permitting it to be filled with liquid nitrogen, covering the grids. Grids are transferred one at a time from the “grid fridge” to holders in the cartridge. When 6 grids are loaded, the top of the chamber is closed with the retractable cold trap, the trap door is opened (dumping the liquid nitrogen surrounding the specimens) and the chamber is evacuated using a large diaphragm pump followed by a turbo pump and finally an ion pump. The cold trap surrounds each specimen on three sides, blocking any path between specimens. A computer controls specimen and trap temperatures through a simple USB interface. The control program rapidly warms the specimens to –160C, slowly warms the samples from –160C to –80C overnight, then rapidly warms to room temperature.
Upon completion of the freeze-dry cycle the cold trap is raised, the rack containing specimens is withdrawn into its shell (and sealed) and the exchange cartridge (with specimens under vacuum) is carried to the STEM, bolted to the specimen changer air lock, indexed to the desired specimen and that specimen is inserted into the specimen stage.
The need to maintain the samples under vacuum after freeze-drying arises because of the tendency for protein and salt residue to be very hygroscopic. When exposed to air, they both absorb water vapor, then effectively air dry when evacuated for electron imaging, causing all the artifacts associated with air-drying. Furthermore, any salt or denatured protein that collects between grains of ice is dispersed and less obvious on the specimen. (The initially amorphous ice tends to form crystals when warmed above –100C, while most of the sublimation occurs above –90C.)
Final specimen quality is a function of both local factors (differing from area to area on the grid) as well as global factors. Therefore STEM operators are careful to collect data from several widely separated areas. Possible local factors include: carbon substrate variation, mixing kinetics of small, injected volumes, diffusion rates and attachment fractions of various species, surface charge or film breakage on the substrate and thickness variation in the ice (prior to freeze drying). Global factors usually relate to the initial sample solution such as purity, concentration, presence of fragments or denatured protein and buffer problems.
Last Modified: June 12, 2009