MEMS Metrology

1
MEMS Metrology

• Jeremy Donaldson
• HP Imaging and Printing Group
• ISMI AMAG/OMAG 9 /2006

2
Todays topics
HP MEMS overview This stuff is cool A few case
studies Why current techniques are falling short
3
Inkjet printing system

• Key components
• Ink
• Writing System
• Ink Delivery
• Carriage
• Paper path
• Interface
• Media
• Printhead

Office
Home
Professional
Retail
4
HP Scalable Printing Technology (SPT)HP RD
invested more than 4 years and 1.4B in SPT
Enterprise
Retail application
Industrial printing
Graphics
Home printing
SMB office
5
HP Innovationdefining an inkjet Moores Law
HP inkjet performance has doubled every 18 months
for the last 20 years
Source HP R D
SPT
1 billion
Printheaddrops persecond
100 million
10 million
1 million
1 thousand
1980
1985
1990
1995
2000
2005
2010
2015
Product of (drop rate) X (number of nozzles) on
a single silicon chip. Used for comparison
purposes, not necessarily achieved in practical
print modes.
6
HP Scalable Printing Technology

• Full lithographic process
• Can vary critical dimensions easily and quickly
  for different products
• All leverage same fundamental process

7
HP Thermal Inkjet Technology
Microscopic view of a drop generator(with nozzle
plate removed)

• Microscopic view under stroboscopic illumination
• Up to 36 000 vapor bubble cycles per second

8
Making an SPT Inkjet Printhead (1)thin-films and
heaters

• silicon wafer
• grow oxide layer
• deposit conductor film
• etch window for resistor
• deposit resistor film
• etch conductor and resistor
• deposit dielectric film
• deposit tantalum film

9
Making an SPT Inkjet Printhead (2)drop generator
chambers, channels, and pillars

• apply photosensitiveepoxy layer
• use a negative exposure mask
• expose
• chemically developunexposed areas

10
Making an SPT Inkjet Printhead (3)orifices

• fill chambers with waxto stabilize structures
• add photosensitive epoxylayer
• use a negativeexposure mask
• expose
• chemically developunexposed areas
• remove wax

SPT a monolithic printheadassembled with
submicron precision
11
HP Scalable Printing Technology a breakthrough
in printhead architecture

• entire printhead created photolithographically
• precision alignment of chamber and nozzles to
  heater improves drop placement accuracy
• 1200 nozzles/inch across the ink feedslot
• high nozzle density for fast printing

Orifice (nozzle)
Top Plate
ink flow
Pillar
Chamber
thin films
Silicon Substrate
Photo credits HP R D
12
(No Transcript)
13
Put it in perspective

• 5ng drop
• 10m/s
• 1-2mm pen to paper spacing
• 4800dpi grid (5.25um)
• 0.13deg included angle

14
HP Scalable Printing Technology detail of drop
generator and particle filters

• small feature size for high nozzle density and
  high refill speed
• filters help guarantee reliability over the life
  of the printer

Mesh filter(not shown)1st filter
Example a red blood cell 8um
15
HP MEMS metrology needs

• Inkjet is about ejecting drops this is largely
  defined by the firing chamber and the nozzle
• Key parameters dimension in X, Y, and Z AND
  angles
• Frequently re-entrant, very deep, and/or
  completely enclosed
• Range from the small (/- .5um out of 25um) to
  the relatively large (/- 25um out of 625um)

20um
700um
16
Some specific needs - Bores

• Bores are created in place
• Re-entrant feature
• Taper angles are very critical needs down to
  0.1deg accuracy
• Entrance and exit diameters
• Layer thicknesses top layer, air gap inside
• Charging is always an issue
• Non destructive is highly desirable

15-20um
15-20um
5-15deg
15-20um
17
Whats been tried, and issues

• Confocal Laser Scanning Microscopy
• Used successfully for previous technologies in
  reflected mode
• Can take slices at different Z levels to build
  a model of the structure
• When dealing with enclosed features, need to use
  fluorescent mode
• Resolution plunges
• Relies on either natural fluorescence or on
  additives to films so they fluoresce consistently
• Very tight angle specs pushed the limit of even
  the modified films
• Needed oil or water immersion to improve accuracy
• Slow

18
Whats been tried, and issues

• Optical Microscopy
• Used successfully for top level focus XY
  dimensions
• Can image through transparent films for substrate
  measurements and registration measurements
• Limitations
• Thick film focus offsets
• Cant provide accurate information about bore
  angles
• Even some question about exactly what is being
  measured (just what is a thick black line
  really?)
• Have seen impacts of things like film stress
  changing RI, reflections, etc.

19
Whats been tried, and issues

• Focus Ion Beam / SEM
• Can cut through to see features
• In theory can provide accurate measurements
• Limitations
• Destructive
• Accuracy large feature size and charging
  effects cause semi-systematic errors in
  measurement
• Slow
• Very sensitive to tool setup

20
Some specific needs - Slots

• Slots that stop partway through
• Need to know X, Y, Theta
• Depth at center of slot as well as some measure
  of the degree of flatness across the bottom
• Slot bottom roughness is high and aspect ratio is
  high.
• Again, non destructive is highly desirable

Laser
Wet Etch
Hybrid
21
Whats been tried, and issues

• Techniques used
• Front side /back side camera
• White light interferometer
• Challenges for HPs applications
• Current algorithm to determine a trench bottom
  is not robust
• Leads to poor depth and X_align gauge RR
• Current algorithm to determine a trench end is
  not robust
• Leads to many missed or incorrect trench end
  location
• Tool integrators choice of optics is not robust
  for our application
• Wafer bow (up to 600um on 200mm wafer) difficult
  for tool hardware and software to handle
• Tools front to back alignment correlation drifts
  due to thermal, vibration and stage move errors
  causing drift in tools alignment calibration

22
Some specific needs Epoxy posts

• Post height diameter would be pretty
  straightforward
• But of course, we need specific information about
  the angle of the posts in at least 2 axis
• And, of course, nondestrucive and fast

23
Whats been tried, and issues

• FIB/SEM
• Can provide quantitative 2D information
• But what we really need is quantiative 3D
  information
• Slow
• Destructive

24
Whats been tried, and issues

• AFM and AFM with flared tip in lateral tapping
  mode
• Standard AFM cant resolve very steep sidewalls
• Flared tip AFM only gets first 500nm of the posts
• Speed

25
MEMS metrology - mitigation strategy

• Some ways to address the issues with MEMS
  metrology
• Design process such that critical features can be
  measured
• Design sensitivity out of features that are
  difficult/impossible to measure
• ISMI projects

26
MEMS Metrology Summary

• A few things that make MEMS challenging
• Combination of size and accuracy
• Compared to ICs theyre big however
• Precision/Accuracy needs are tight
• Mixed materials
• Metals
• Dielectrics
• Polymers / Organics
• 3D
• Out of plane
• High DOF needed
• Enclosed/obstructed cavities

27
Request to SEMATECH

• MEMS is growing already several killer apps
  (HP inkjet, TI DMD, Analog Devices accels, Si
  microphones) there are going to be more more
• SEMATECH is pushing smaller but MEMS metrology
  is for larger features.
• Consider what can be rolled back up the scale
  
• Techniques
• Algorithms
• Leveraged platforms
• The SEMATECH rigor would be welcome

28
The wish list

• Non destructive
• Immune to charging (material properties)
• High resolution AND large field of view
• Ability to measure reentrant / enclosed volumes
• Ability to measure out of plane (at least
  orthogonal, ideally any orientation)
• Ability to measure rough (ums of roughness)
  surfaces
• High throughput / automation

29

• Backup Slides

30
Concentricity Requirements

• Bore concentricity, ?, correlates with drop
  trajectory.
• Small angles matter.
• All products have spent considerable resources
  improving trajectory performance.