Transcript High Resolution 10-27-2008

Lab report 4
•Very important: if you raise the sample to
the 10 mm line during pumpdown, the
working distance will be approximately 10
mm. (How do you know what the WD is?)
•There is no “elephant” (pink or otherwise)
for the Quanta.
•Some people still aren’t getting the hint
about attaching a copy of their lab notes
•Nobody reminded me to upload images!
Nanowires, 1 MP
Nanowires, 512x442
High resolution imaging
• Why?
– To see really tiny stuff!
• Soot particles (Combustion research)
• Gold nanoparticles and nanorods (Jennifer
Shumaker-Parry)
– To get really sharp images of “fairly small” stuff
• Gold nanoparticles and nanorods (Marc Porter…)
– Same considerations should apply to e-beam
lithography
Tin balls 2.4 Mx!
Soot on TEM Grid
High resolution imaging
• How?
– FEG Why?
• dp2 = 4ip/( π2 * β * αp2)
– To see small stuff, you need a small spot.
– Small spot requires
• Low probe current (small “spot size”)
• High brightness
• Large aperture angle
– αp = dA/2*WD
– Short working distance
– Large aperture? We’re still arguing over that one, but
nobody believes it!
Contributions to actual beam diameter
Optimum convergence angle
• The previous equation can be
differentiated and the derivative set equal
to zero
• Result is quadratic in α4
Optimum convergence angle
• If chromatic aberrations can be neglected:
– αopt = [2/3Cs2)(8ip/βπ2 + 0.72 λ2)]1/8
– At 10 kV, λ = 0.01 nm, 100 pA, and β = 108 A/cm2sr
Cs = 2 mm, then αopt = 6.4*10-3 rad
– αp = dA/2*WD
• If dA = 30 microns: WD = 2.4 mm
– At 1 kv λ increases by a factor of 10
• αp increases to 9 mrad
• WD decreases to 1.7 mm
• Much shorter for W-filament!
• Note that αopt was 4 mrad for W-filament
(Cs = 10 mm), 20 kV in Fig 2.22.
High resolution imaging
• What’s the point of all this math for high
resolution imaging?
• Short WD is important!
• ETD is out of the game
• Other options
– TLD
– vCD
– Helix
• Immersion mode increases resolution
about another order of magnitude
High resolution imaging
• Minimize emi (Ian’s lecture of last Friday!)
• Minimize floor vibrations
– Each lab in INSCC is on an individually
poured slab of concrete
•
•
•
•
Minimize acoustic vibrations (audio noise)
Minimize temperature fluctuations
Try “Mains lock”
Replace carbon tape with silver or carbon
paste
• Use single sample mount
Minimize Acoustic vibrations!
When do we run out of
resolution?
SE1 have a range of a few nm and
create the ‘edge bright line’ effect
DCJ - High Resolution
As a result when the feature size is
close to the SE escape range l the
object is not resolved. This occurs
at ~ 5-10nm for low Z materials
14
In other samples...
 When an object gets
small enough to be
comparable with the
SE1 generation volume
then it becomes bright
all over and the
defining edges
disappear.
 For low Z, low density
materials this can
happen at a scale of
5-10nm
DCJ - High Resolution
edge brightness
no edges
Carbon nanotubes
5nm and 10nm wide
15
SE image of Single Wall NanoTube
DCJ - High Resolution
16
..but how wide is this
nanotube?
The image does not have defined edges its width is indeterminate and equal to
If these are the
lcarbon
If these are the
edges then
width = 5.8nm
edges then width
= 1.5nm
?
lC =5.5nm
DCJ - High Resolution
17
Bypassing the SE1 limit
 Metals have lower l than
carbon, and a higher SE
yield
 A thin metal film on a low
Z, low density sample
localizes all SE production
within itself. The resolution
now is a function of the
film thickness only
 Works even with very thin
metal films (a few atoms
thick)
 We can exploit this effect
to give interpretable
contrast beyond the
theoretical limit
DCJ - High Resolution
High SE
yield
Low SE
yield
width ~ film thicknes
even when < l
18
Mass thickness contrast
 The SE1 yield varies
with the thickness of
the metal
 SE1 yield reaches the
bulk value at a film
thickness equal to about
3l
 The conformation of the
film to surface
topography provides
contrast
DCJ - High Resolution
S
E
bulk value
Y
i
e
l
d
mass thickness
variation
1nm
2nm
3nm
Film thickness
19
Metal builds contrast
 The SE localization in
the film provides edge
definition
 The mass thickness
effect gives extra
contrast enhancement
 The feature is now
‘resolved’ since its
size and shape are
visible
2nm metal film
5nm low Z
object
S
E
Beam
position
SE profile without metal
SE profile with metal film
DCJ - High Resolution
20
Cr coatings
 Cr films are smooth and
without structure even as
thin as 1nm
 The mass thickness
contrast resolves edges and
make the detail visible down
to a nanometer scale
 The high SE yield of the Cr
improves the S/N ratio
 However these coatings are
not stable - so use Cr
coated samples immediately
after they have been made
Phage
T4 T4
Phage
+ Cr
DCJ - High
coated
withResolution
Cr
courtesy of Martin Müller
and Rene Herrmann, ETH Zürich
21
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 particulate coatings are useful
because of their higher SE yields and better S/N
ratios
 Above 100kx can use chromium or titanium or
particulate coatings of W,Pt or Ir
 Carbon is a contaminant not a coating
DCJ - High Resolution
22
Helix Detector
•
•
•
•
•
•
•
Looks just like the LowVac detector
Mounts just like the LowVac detector
Costs $28.5k
“You break it, you bought it.”
Runs in Immersion Mode, like the TLD
Gives super images in LowVac
Aperture is 62 microns
Helix Detector page
Helix detector settings
• 1- why and when should Helix detector voltage be changed?
• Helix detector voltage should be optimized after every parameter
change. This just means, from the previous Helix sweet spot,
needing to bump it up or down a 1% or 2% after changing a
parameter like water, FWD, beam current, kV, dwell.
• I believe Helix needs to be running at full current, just prior to arcing,
at all times. A typical Helix session may run like this; 78.5% voltage
= arcing, 77.5% voltage = no arcing and good signal contrast, 76.5%
voltage = no arcing and poor contrast. I don’t expect you to see
identical numbers as these but I’d expect you to experience a similar
“narrow window” of good zone.
• For me, Helix sweet spots are 1%-1.5% off saturation, 3.2-mm
FWD, 4kV- 7kV, as much as 30% less water than what was needed
for LVD.
• Reducing Spot size just means increasing dwell times but I typically
work between spots 2.0 and 3.0
Gold on glass, LowVac, Helix
Detector
vCD Detector
•
•
•
•
•
Backscatter detector
Mounts on pole piece
Has small aperture; will work in LowVac
Works with beam deceleration
Works in immersion mode
Landing Energy Page
Gold Nanoparticles, TLD
Gold Nanoparticles, vCD
Worm cross-section, vCD
Virus, vCD