Field Induced Damage Mechanisms In Reticles - Sematech Presentation Dec 2003

Presentation (PDF Available) · December 2003with 2 Reads
Sematech ESD Symposium
Abstract
This is the first presentation I gave to describe my discovery of EFM and to explain how I thought the damage mechanism worked.
Microtome
Reticle protection is our business
Gavin Rider
Vice President, Technology and Development
Microtome
Reticle protection is our business
Microtome Precision 2004. CONFIDENTIAL 2
Reticle ESD 2:
Investigating the nature of
field induced reticle damage
Microtome
Reticle protection is our business
Microtome Precision 2004. CONFIDENTIAL 3
Experiments into field
induced ESD damage
Microtome
Reticle protection is our business
Microtome Precision 2004. CONFIDENTIAL 4
Sematech study 1
V
~3 cm
Insulating panel
Canary test reticle
1x Ramped
high voltage
Electrode simulating
an operator’s finger
There is no conductive ESD – the reticle remains neutral.
Microtome
Reticle protection is our business
Microtome Precision 2004. CONFIDENTIAL 5
Guard Ring
1
2
3
4
5
32
33
34
35
~1.5μm
“Flock of birds”
“Bird”
“Beak” The “Canary” test reticle is patented by Dupont Photomasks
#35
#1
Canary ESD test reticle
Microtome
Reticle protection is our business
Microtome Precision 2004. CONFIDENTIAL 6
uWith a charge located centrally
above the reticle we might expect to
see a radially symmetric damage
signature, but the flocks are not
symmetric to the field lines
uThe “R” flocks closest to the centre
of each side would sense the
greatest potential gradients
uThe high birds in each flock are
aligned along an equipotential so
would show little or no damage
uThe outermost flocks have little field
gradient along the entire line of birds
…so this is the predicted damage
pattern.
Field induced ESD in a Canary
Representation of the component of the electric field in the plane of the reticle
Microtome
Reticle protection is our business
Microtome Precision 2004. CONFIDENTIAL 7
Damage pattern predicted for a radial
field within the Canary reticle
A R20
A R15
A R3
A R10
A R5
A R1
A L1
B R1
Reticle 58638
Report MA00357
Single 9kV voltage
ramp on electrode
above reticle
Results from Sematech study 1
This shows the type of damage that occurs at different induced potential differences.
Microtome
Reticle protection is our business
Microtome Precision 2004. CONFIDENTIAL 8
AFM scan of reticle 58638 A_R5 gap
Vertical scale x10
The A_R5 gap is
damaged by two
mechanisms:
1) Pairs of bumps on the
top surface and
roughened upper line
edges are the damage
sites from air discharges
across the gap from the
corners of the bird beak
2) The growth of a smooth
protrusion on the surface
between the lines is
caused by another
process. It appears to be
due to chrome migration
Results from Sematech study 1
Microtome
Reticle protection is our business
Microtome Precision 2004. CONFIDENTIAL 9
The A_R10 gap shows no signs of air
discharge, only growth of a protrusion
between the lines
1) There are no pairs of bumps on the top
surface and there is little abnormal
topography on the line edges
2) A smooth protrusion is growing from the
base of the “bird body” towards the line
end of the “bird beak”
AFM image of reticle 58638 A_R10 gap
4) The protrusion is growing only in one
direction. This shows that this process is
sensitive to the polarity of the electric
field between the features.
3) This is forming centrally in the gap,
showing that its growth is not related to
the sharp corners at the end of the “bird
beak”
Results from Sematech study 1
Microtome
Reticle protection is our business
Microtome Precision 2004. CONFIDENTIAL 10
The bridge growth starts at the base of the line. As the potential difference across the gap
increases (towards the guard ring) the extent of the growth increases.
Vertical scale x10.
AFM surface profile measurements of reticle 58638, report MA00357
Centre of reticle (top left) moving towards the guard ring (lower left).
The building of the bridge
Microtome
Reticle protection is our business
Microtome Precision 2004. CONFIDENTIAL 11
How the bridge builds up between the lines
Vertical scale x10.
Surface
deposit seen
Growth starts
at the bottom
of the side wall
The building of the bridge
Microtome
Reticle protection is our business
Microtome Precision 2004. CONFIDENTIAL 12
Sematech study 2
V
~3 cm
Insulating panel
Test reticle
20x Ramped
high voltage
Electrode simulating
an operator’s finger
RF antenna
There is no conductive ESD – the reticle remains neutral.
Microtome
Reticle protection is our business
Microtome Precision 2004. CONFIDENTIAL 13
Reticle 58197. 20 x 10kV above reticle
3) There is a build-up of material along
the edges of the lines in the gap and
back along the bird beak.
2) A bridge is forming centrally in the
gap, with the protrusions progressing
equally from both sides.
1) The equal protrusions on both sides
reflect the bidirectional induced
potentials, in accordance with the
EMI event data.
AFM image of reticle 58197 B_R23 gap
Results from Sematech study 2
4) Some beading is seen on the quartz
surface between the protrusions.
Microtome
Reticle protection is our business
Microtome Precision 2004. CONFIDENTIAL 14
AFM scan of the B_R23 gap
Vertical scale x10
Perspective view
shows evidence of
material migration:
1) No bumps on the top
surface and no
abnormal upper line
edges, showing that
no air discharge has
taken place.
2) Protrusions are
growing from the
base of both lines,
with surface beading
visible on the quartz
in between.
Results from Sematech study 2
Microtome
Reticle protection is our business
Microtome Precision 2004. CONFIDENTIAL 15
Another Canary reticle gap
shows the lines beginning to
spread out:
1) The bumps on the top surface
indicate that some air discharge
has taken place.
2) There are no protrusions growing
in the gap, but the lines on either
side of the gap are spreading out
across the quartz.
Other field induced damage
3) There is some faint beading
visible on the surface of the
quartz.
Microtome
Reticle protection is our business
Microtome Precision 2004. CONFIDENTIAL 16
Field induced damage summary
uAt the edge of the reticle near the guard ring where
the induced voltages are highest we see evidence
of air discharges. (“ESD”)
uCloser to the edge of the reticle as the induced
voltages become higher, we see protrusions
forming between lines which eventually form
bridges across the gap. (“EFM type 2”)
uAt sites near the centre of the reticle where the
induced voltages are lowest we see the edges of
the lines starting to spread out onto the quartz.
(“EFM type 1”)
Microtome
Reticle protection is our business
Microtome Precision 2004. CONFIDENTIAL 17
Characteristics of the
three different
damage mechanisms
ESD
Microtome
Reticle protection is our business
Microtome Precision 2004. CONFIDENTIAL 18
Point-to-line ESD damage
uIn reticles exposed to electric fields, sites that suffer ESD damage
are often those with the highest local field strength, such as
corners and line ends.
Microtome
Reticle protection is our business
Microtome Precision 2004. CONFIDENTIAL 19
uESD damage is also seen between lines that are running parallel
to each other.
Line-to-line ESD
uSparks are thought to be triggered by local field amplification, from
random air ionization events that occur naturally.
Transmitted image Integrated imageReflected image
uThe potential difference between such lines is identical along the
entire length of the lines, so why is the discharge localised?
Microtome
Reticle protection is our business
Microtome Precision 2004. CONFIDENTIAL 20
ESD overview
uIf the induced potentials are high enough, a discharge may be started by
field emission of electrons from the corners of a negatively biased line.
ESD in reticles is generally associated with field concentration points on
corners and line edges.
uDischarges can also happen randomly along line edges as a result of field
amplification effects caused by natural air ionization events.
uIn all cases, the damage caused by air discharge is distinctive and easily
identified. It is highly localised and is usually severe, involving ablation of
antireflection coatings or melting / vaporization of the chrome.
uAt the lowest voltage that is sufficient to trigger a discharge, the damage
can be relatively minor and no material ablation occurs. It is thought that
these minor air discharges in reticles do not involve air breakdown.
A full presentation covering the analysis behind
these conclusions is available on request
uAt voltages below the onset of ESD, other phenomena are seen…
Microtome
Reticle protection is our business
Microtome Precision 2004. CONFIDENTIAL 21
Characteristics of the
three different
damage mechanisms
EFM type 1
Microtome
Reticle protection is our business
Microtome Precision 2004. CONFIDENTIAL 22
uThese three gaps were all damaged in the same experiment. The gap
potential is different in each case due to the position of the features on
the reticle.
uThe outward extent of the line spreading is the same wherever it has
occurred (not including the initial stage of protrusion growth).
uThe line spreading appears to be dependent on factors that were the
same for all these gaps, such as the temperature and the time for which
the field was applied.
ab c
CD Degradation by EFM type 1
Va< Vb< Vc
Microtome
Reticle protection is our business
Microtome Precision 2004. CONFIDENTIAL 23
uThe locations where line broadening takes place can be related to the field pattern
and its strength at the point where it intersects the metal surface:
ab c
Field line configuration
uThe field line configuration is the same for all the gaps, but the voltage gradient
along each field line depends on the voltage induced across each gap.
uA field line with a particular field strength Emwould intersect the metal lines at
different positions, depending on the gap voltage. As the voltage is increased the
point of intersection with the chrome of the field line with strength Emmoves
outwards along the bird body and back along the sides of the bird beak.
uWherever the field at the metal surface is >Emmaterial migration is seen to occur
at a fixed rate, indicating that the field enables the migration but it does not drive it.
EmEmEm
Va< Vb< Vc
CD Degradation by EFM type 1
Microtome
Reticle protection is our business
Microtome Precision 2004. CONFIDENTIAL 24
uThese images show that
the line spreading is greater
on the straight edges than
at the line corners.
uLocal field strength is
greatest at line corners, so
local field strength is not
the driving force for the
material migration.
CD Degradation by EFM type 1
Microtome
Reticle protection is our business
Microtome Precision 2004. CONFIDENTIAL 25
uAt the corner cell there is also a unit of
sidewall of length l. In time tthe flux
Ffrom this part of the wall can also fill
an area A .
lA
uThe sidewall is divided up into units
of length l.
F
uHence the migration progresses further from the sidewall on the straight
line edges than at the corners.
Let the outward flux of
atoms per unit length be F. In time t
this flux can cover a surface area A
with a monolayer of atoms.
uThe net sideways flux across the cell
boundaries along the straight edge is
zero, but in the cells next to the
corner there is some net sideways flux. This reduces the coverage in
these cells and adds slightly to the corner cell coverage.
Model for EFM type 1
Plan view of a line corner
Microtome
Reticle protection is our business
Microtome Precision 2004. CONFIDENTIAL 26
Characteristics of the
three different
damage mechanisms
EFM type 2
Microtome
Reticle protection is our business
Microtome Precision 2004. CONFIDENTIAL 27
uSome surface beading is generally visible on the quartz when field
induced material migration has occurred. The amount of the beading
would be related to the surface density of the metal atoms on the quartz:
uAt low surface migration currents
the atoms would be far apart on
the surface and would not
interact greatly with each other.
uAt high currents there would be
an increased chance that the
migrating metal atoms would
meet others on the surface and
coalesce into globular beads
due to surface tension forces.
Surface beading
uA distinct surface track like this
shows the start of EFM type 2
Microtome
Reticle protection is our business
Microtome Precision 2004. CONFIDENTIAL 28
uThe path of the metal migration that leads to the growth of the protrusion
is sensitive to the direction of the external electric field. When the field is
aligned at a significant angle to the axis of the conductive features the
protrusion grows at an angle across the gap. (These images are from
gaps near the centre of the reticle, where the field direction is obliquely
across the line of the flock).
uThis indicates that the migrating material that develops the surface tracks
and ultimately the protrusion is probably charged.
Gap 15 Gap 15 Gap 30
Protrusion growth by EFM type 2
Microtome
Reticle protection is our business
Microtome Precision 2004. CONFIDENTIAL 29
uThe protrusion grows where the average
field strength across the gap is highest.
uThe growth direction is related to field direction.
uThis suggests a protrusion growth mechanism
that behaves like an electric current:
uThe extent of growth is dependent on the
potential (or the average field strength) between
the features.
ØIf the mechanism involved momentum transfer to the chrome atoms from
an electron current (electromigration) the atoms would move in the
direction of the electron current. This is not seen. Also a high current
density is required for electromigration and this is not present.
ØThus the mechanism is thought to involve ionic surface conduction, which
would only take place from the positive electrode, assuming the mobile
species are Cr+ions. This matches the growth direction seen in the single
voltage ramp experiment.
Protrusion growth by EFM type 2
Microtome
Reticle protection is our business
Microtome Precision 2004. CONFIDENTIAL 30
Can it really be
chrome migration?
Microtome
Reticle protection is our business
Microtome Precision 2004. CONFIDENTIAL 31
From the scientific literature…
J. Vac. Sci. Technol. B, Vol. 20, No. 3, May/Jun 2002
Surface energy has both
mechanical and
electrostatic components
Microtome
Reticle protection is our business
Microtome Precision 2004. CONFIDENTIAL 32
From the scientific literature…
uThis shows that it is energetically favourable for surface atoms to
migrate under the influence of electric fields and to form protrusions
on solid surfaces. This reduces the total energy of the system.
Under electrostatic stress a
planar surface tends to form
protrusions like this:
Microtome
Reticle protection is our business
Microtome Precision 2004. CONFIDENTIAL 33
Wherever the chrome migrates the contact angle with the quartz is about 20° (note
the 10x vertical magnification in the AFM scans). This shallow contact angle shows
that it is energetically favourable for chrome to spread on quartz. (The same
physical property is responsible for the strong adhesion of chrome films to glass).
Metal migration characteristics
Topography scans from different reticles exhibiting material migration
Actual profile Actual profile Actual profile
Actual profile Actual profile Actual profile
Microtome
Reticle protection is our business
Microtome Precision 2004. CONFIDENTIAL 34
u1. Line spreading that degrades CD takes place at a
seemingly constant rate once a certain amount of
electric field is present on the line edge.
u2. Protrusion growth that leads to gap bridging is
sensitive to both field direction and the average field
strength in the gap.
èThis type of migration does not seem to be affected by the
field direction or strength - it is proposed that it involves the
thermal diffusion of neutral atoms.
èThese characteristics suggest the conduction of metal ions
across the quartz surface.
There are two patterns of metal migration:
Metal migration summary
Microtome
Reticle protection is our business
Microtome Precision 2004. CONFIDENTIAL 35
u1. CD degradation occurs at the lowest levels of induced voltage.
There is a threshold below which metal migration does not take place -
the level of this is not known but it is very low by ESD standards. It is also
not known whether this migration process will continue indefinitely or
reach a point of equilibrium, perhaps related to the chrome film thickness.
u2. Gap bridging occurs at medium levels of induced voltage.
Once a gap is bridged currents flow back and forth continuously when the
fields around the reticle change, causing localised Joule heating. If the
currents are low this heating progressively oxidises the surrounding
chrome lines creating “chrome foam”. If the currents are high enough the
heating can vaporise the bridge completely.
There are three physically different forms of damage that take place
depending on the amount of reticle exposure to electric fields:
Conclusions
u3. Air Discharge occurs at higher levels of induced voltage.
The lowest-energy discharges land on the upper surface of the chrome
and so can damage the anti-reflective coating. Higher-energy discharges
vaporise metal from the line edges and cause “mouse bites” (these are
the most commonly known features of ESD damage).
Microtome
Reticle protection is our business
Microtome Precision 2004. CONFIDENTIAL 36
Implications for advanced masks
uOPC features
Non-printing optical-proximity-correction features on binary
masks could be affected:
if the OPC features broaden, causing too much light to be
blocked, the features themselves might become printable.
if the migrating chrome becomes oxidised and hence the
features become increasingly transparent, this would reduce the
power of the OPC features and their effectiveness would be lost.
uLine-end serifs
uReticle lines that are written with serifs to make the line-ends
print correctly could be badly degraded by line-shape changes
of the kind observed on the Canary reticles.
uPSMs
In phase shift masks any migrating material deposited onto the
quartz surface between features would affect the phase of the
transmitted light and degrade the reticle.
Microtome
Reticle protection is our business
Microtome Precision 2004. CONFIDENTIAL 37
Main Conclusion
It should not be considered
safe to expose ANY reticles to
ANY level of electric field.
Microtome
Reticle protection is our business
Microtome Precision 2004. CONFIDENTIAL 38
Acknowledgements
Andy Rudack, Asmita Shah and Susan Elder of
SEMATECH for the AFM data
Larry Levit of Ion Systems for conducting the
experiments and for valuable discussions
Many thanks for supply of the electrostatic modelling software
and for excellent technical assistance:
Vector Fields Ltd., 24 Bankside, Kidlington, Oxford OX5 1JE, UK
Tel: +44 (0)1865 854999 Fax: +44 (0)1865 370277
info@vectorfields.co.uk www.vectorfields.com
Microtome Precision, Inc
4805 Northpark Drive
Colorado Springs, Colorado 80918 USA
719-598-8831 719-598-3316 fax
e-mail: microtome@aol.com
website: www.microtome.com
Microtome
Reticle protection is our business
This research hasn't been cited in any other publications.
This research doesn't cite any other publications.