Andrew Pomfret

1
Andrew Pomfret

• Flatness of Carbon Films for Electron Microscopy

2
High-resolution electron crystallography at high
tilt angles can't be realized by the intrinsic
stiffness of protein crystals alone-- needs to
sit on a flat carbon support -- whose flatness
is dependent on the carbon film, EM grid, and
preparation techniqueCARBON SUPPORT FILMS
Used as support films for samples on EM grids as
provides mechanical stability, high transparency
for electrons, and good conductive
properties. BUT, for 2D crystallography, lack of
crystal flatness is a major cause of limited
resolution, and carbon non-flatness has been
shown as a major cause of this.
3
CARBON SUPPORT FILMS FOR 2D ELECTRON
CRYSTALLOGRAPHY
Our Problem- Limited resolution diffraction
perpendicular to tilt axis due to lack of
specimen flatness on holey carbon grids
Flat Carbon Films Atomically flat mica-surface
produces very flat carbon films, but need to blot
from front-- blots off large flat 2D crystals
Holey Carbon Films Good for blotting from behind
to suck large 2D crystals onto grids, but holey
plastic very uneven
4
PREPARATION OF FLAT CARBON FILMS
5
SOURCE OF CARBON FILMS
However, people at the 2D crystal workshop all
seemed to like high purity carbon (gt99.9999) and
very high vacuum (lt2x10-6 Torr vs. 0.9-2x10-6
Torr by Han et al.)
6
CARBON EVAPORATION (Bradley 1953)-EFFECT OF
SPARKING AND LAYERING
(Gyobu 2006)
Different papers are divided about whether doing
many small layers of carbon on top of each other
(multiple evaporation) is better than doing a
single-burst evaporation layer (more thin
layers are reported by some as flatter, whilst
others at worst say there is no difference, so
probably best to layer!).
7
CARBON EVAPORATION-- EFFECT OF LAYERING
A paper by Butt et al (with Da-Neng Wang
co-author 1991) claims multiple evaporation is
better, and showed that the mica-touching carbon
side is 3-9 x flatter than the upper side by AFM
(see figure)
8
NATURE OF CARBON FILMS ON EM GRIDS
9
QUALITY OF EM GRIDS
Also, less wrinkling of carbon has been found by
SEM when floating carbon onto the dull side of Cu
grids (Brisson 1996)
10
Various Answers (Glaeser 1992)-Carbon
evaporated on mica (the side facing the mica)
approaches the flatness of the mica by AFM
Upper surface of carbon evaporated on mica is
up to 8.6 deg/um by STMCarbon evaporated on
single crystal graphite up to 34.4 deg/um by STM
HOW FLAT IS CARBON?
NOTE- For high resolution diffraction at high
tilt-- specimen must remain flat to a tolerance
of better than 1 degree (Glaeser 1993) or
possibly 0.2 degrees per um (Glaeser 1991).
11
AFM of Carbon Sandwich Surface(Koning 2003)-
Maximum curvature of 0.2 deg/um
12
EFFECT OF CRYO-CRINKLING ON CARBON
FLATNESSTheoretical calculations of 0.023
deg/um, but uncertain of validity of calculation
(Glaeser 1992).However, seems to be a large
difference in obtained diffraction using Mo grids
vs Cu grids as Mo thermal expansion coefficient
closer to that of the carbon film (both appear
flat at RT) and SEM images clearly show crinkling
(Booy Pawley 1992). They also found
mica-floated-carbon deposited on holey carbon
support film was flatter across the holes,
especially when further from the Cu grid bars.
13
NATURE OF CARBON FILMS ON EM GRIDS
14
FOUR IDEAS TO MAKE SPECIMENS FLAT (Glaeser
1992)(excluding carbon sandwich, covered in
later slides)
Purple Membrane
15
METHODS OF LOADING SAMPLES ONTO SINGLE CARBON
FILMS ON EM GRIDS
16
METHODS OF LOADING SAMPLES ONTO SINGLE CARBON
FILMS ON EM GRIDS
INTERESTING- To avoid problems of crystal
curvature on holey carbon, Kuhlbrandt et al
(1999) grew crystals directly on
flat-carbon-coated grids placed vertically on
drops of their crystallization (as their crystals
formed at the air-water interface).
17
METHODS OF LOADING SAMPLES ONTO SINGLE CARBON
FILMS ON EM GRIDS
18
METHODS OF LOADING SAMPLES ONTO SINGLE CARBON
FILMS ON EM GRIDS
19
MAIN PROPOSED SOLUTION TO THESE PROBLEMS
20
(No Transcript)
21
(No Transcript)
22
ADVANTAGES-1) Sample (i.e. large 2D crystals)
trapped between two layers of carbon, so not
blotted off easily2) Reduces beam-induced image
shift 3) Apparently more reproducible/consistent
ice thickness, according to some papers and
people.4) Perhaps two flat layers would help
hold a 2D crystal more rigidly flatDISADVANTAGES
-1) More tricky to carry out (although not
necessarily with the alternate method)2)
Difficult controlling ice-thickness-- variable
blotting and air-drying time
23
(No Transcript)
24
(No Transcript)
25
2D protein crystals trapped on carbon sandwich
grid, trapped between two flat carbon sheets with
nice flat even ice (Koning 2003).(reproducibly
thin and suitable for cryo-electron microscopy)
26
CONCLUSIONS ON CARBON FILM PROPERTIES

• Glaeser (1993) conclusion-- ...specimen
  flatness... depends very much on the surface
  properties of carbon film. Unfortunately, these
  surface properties can vary greatly from one
  day's preparation of carbon film to another, from
  one laboratory to another, and over prolonged
  aging of the same preparation. Even so, extremely
  flat well preserved specimens can be obtained
  after some trial and error...

27
CARBON GRID PLANS FOR OUR 2D CRYSTALS

• 1) Have tried back-injection method with
  face-blotting and lost the large 2D crystals,
  but have found luck with blotting 2 edges
  simultaneously (by dehydration)-- need to test
  evenness of ice.
• 2) Have had difficulties with ice-evenness due to
  hydrophobicity of carbon film-- can't glow
  discharge mica carbon, but can include
  pre-wetting of the carbon grids with
  poly-L-lysene or BSA (both didn't seem to affect
  the adherence of large 2D crystals much, but may
  help with ice evenness)
• 3) Try the much-hyped carbon-sandwich method
  that seems to be a lot of people's ultimate
  solution to a lot of problems with electron
  crystallography
• 4) EMS claims to have ultra-flat holey carbon
  grids (C-flat grids) made without plastic,
  making it the ideal specimen support to achieve
  high resolution data in cryo-TEM. I have been in
  contact with EMS to get some sample grids, but
  they are still making/testing their new
  templates and have promised to let me know as
  soon as they know how their manufacturing of the
  new grids turns out.

28
EMS ULTRA-FLAT C-FLAT HOLEY CARBON
GRIDS(definable hole size and distribution--
5/grid)