This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
Ink Jet Deposition of Inorganic Nanoparticle
Materials as a Route to Desktop Fabrication of
Integrated Logic and Micromachinery
by
Sawyer Buckminster Fuller ARCHIVES
Submitted to the Department of Mechanical Engineeringin partial fulfillment of the requirements for the degree of
Author ...............Department of Mechanical Engineering
May 5, 2000
Certified by......Joseph Jacobson
Associate ProfessorThesis Supervisor
A ccepted by .......................................................Ernest C. Cravalho
Chairman, Undergraduate Thesis Comittee
i ' V
Ink Jet Deposition of Inorganic Nanoparticle Materials as a
Route to Desktop Fabrication of Integrated Logic and
Micrormachinery
by
Sawyer Buckminster Fuller
Submitted to the Department of Mechanical Engineeringon May 5, 2000, in partial fulfillment of the
requirements for the degree ofBachelor of Science in Mechanical Engineering
Abstract
In this thesis, ink jet printing is investigated as a means to fabricate sub-millimeterelectrical and mechanical systems such as micro-actuators and integrated circuits.Nano-crytalline atom cluster dispersions in a solvent-based dispersant are used as
the primary building material, the first known ink jet application of such material to
build microelectromechanical structure. The nanoparticle ink offers a means to addi-
tively build devices out of inorganic materials with material properties far superior to
organic materials and comparable to what is normally created in a vacuum. Demon-strated devices include an electrostatic linear drive motor, a 1-bit radiofrequency(RF) remote sensable tag, in-plane and out-of-plane electrothermal actuators, and acapacitor. All processes conducted were in an open atmosphere at plastic-compatibletemperatures, suggesting a potential route to a desktop fab.
Thesis Supervisor: Joseph JacobsonTitle: Associate Professor
Acknowledgments
I'll start off by thanking me, for enduring all this work, which often has entailed long
unpleasant episodes. I'm hoping my future me will be glad I did. I want to thank my
mother, Patrice Engle, and all of my grandparents, Pearl and Mel Fuller and Murry
and Ernest Lauser, for funding my undergraduate education and that wisdom-like
insight into life, even when it looked like I was never going to graduate. Thanks to
Babak Nivi who invited me to work at the Media Lab, to Colyn and Eric for providing
me healthy competition for good results (and a minimum of espionage and sabotage),
to Brian for the AFM imaging, to Joe for being the model of a good leader and for
letting me take on my own project, to Leila for when she helped remind me once
upon a time why life and school were worth working for, to Dad, Kirk Fuller, for the
love of invention, spirit, and hard work, to Henry for the global perspective and for
being my step dad, to Saul for the hotplate, to Jeremy Levitan for MEMS insight
and help, to a lot of incredible people, some of which I have even yet to meet, and to
the times when I've gotten to experience the happiness, warmth, and love of life.
I can't wait to graduate.
This work was funded under the Defense Advanced Research Project Agency
contract DABT63-99-C-0033 and the Media Laboratory's Things That Think con-
As the two towers were built higher, thermal conductivity from the substrate fell,
30
so eventually both were perpetually wet at the top, mushrooming out. The wet-
mushrooming effect eventually caused the two towers to join. As a result of the large
volume of material at the top of the actuator, sintering time was over 30 minutes.
Deflection for this device was negligible.
4.6 Capacitor
A capacitor was built which demonstrated a capacitance of 9±1 pF using a capac-
itance measuring machine. The dielectric insulator was an acetate-based material
designed as a resin for ink jet printing onto nonporous surfaces. The fabrication pro-
cess included ink jet printing a silver bottom plate, depositing the insulator onto a
250*C hotplate with a pipette, and then depositing another layer of conductive silver
on top of the insulator. Layers of the insuating material deposited by ink jet were
shown to insulate, yielding an all ink jet printed capacitor.
Other insulators attempted included photoresists, spin-on-glass, and polyimide.
In all other cases every device fabricated did not insulate. It was postulated that
conduction was the result of cracking, pinhole defects, or the solvent in the ink dis-
solving the insulating material. Ink jet printing of photoresist, spin-on-glass, and
polyimide were demonstrated.
31
Chapter 5
Conclusion
The goal of this project was to show that nanoparticle ink jet fabrication could be
used to build devices of sufficient electrical and mechanical quality and repeatability
to suggest the process could be used as an alternative fabrication system to vacuum-
deposited CMOS and MEMS-type processes at far lower cost.
5.0.1 Cost
The nanoparticle material was purchased at a price of $55 per gram. Using an estimate
of a 10 minute printing time and continuous droplet ejection at 125 Hz, 10% weight
ink, 1 kg/L ink density, and an 80 pL droplet volume, the expected cost of the material
in a simple device such as a linear motor is 3 cents. Considering a standard MEMS
foundary cost of roughly $3000 per square centimeter for a standard process, and a
turn around time of two months, the potential is there for an alternate approach such
as ink jet to offer a revolution in speed and cost.
Considering that all processing temperatures for the devices were below 300'C, in
earth atmosphere, the costs associated with building a device with nanoparticles can
be expected to be far lower than vacuum processes.
32
5.0.2 Performance
The conductivity of the nanoparticle silver ink jet deposited film was within a factor of
two of that of builk silver. That the micromechanical actuator was physically strong
enough to withstand sonication, and further was able to deflect and resume its position
is indication of mechanical strength. In addition, the vertically printed actuator
suggests the ability to print very complex mechanical structure. The results achieved
with nanoparticle material films in this report and others suggests printing to be a
process that can yield results far in excess of those achievable with organic chemistry
means, and of sufficient quality to be used in integrated-circuit type applications.
A concern is device shrinkage. On average shrinkage was 25% with a rapid sinter.
At lower sinter temperatures over extended periods of time, such as 250 C was found
to allow more vertical shrinkage with less cracking. However, it is expected that
more complex devices will require shrinkage compensation (such as the type used for
inkjection molded parts) or an improved material that shrinks less.
The mechanical actuator result was presented to the IEEE Micro-Electromechanical
Systems 2000 conference in Japan in January.[19]
5.0.3 The limits of size
Figure 5-1 is an image of a 5 pm silver droplet generated inadvertently as a result of
a microscopic ink "splash." It shows the potential for much smaller ink droplets to
be used to make devices with feature sizes competetive with lithographic techniques.
5.1 Applications
This work has received a fair amount of interest from companies interested in inex-
pensive tag technology, as well as from groups interested in systems for one-off and
extremely inexpensive integrated circuits. Considering the far lower costs for the
ink jet process in combination with its demonstrated high device quality, ink jet has
much promise for being the process used for a desktop MEMS and or integrated cir-
33
Figure 5-1: Ink droplet at 4000 dpi generated from a splash demonstrates feature size
comparable to those attained by lithography.
cuit fabrication system where cost, device bulk, and turnaround time are significant
factors.
34
Appendix A
Program for Printing Planar
Electrothermal Actuator
The following is a verbatim reproduction of the program used to print the electrother-
mal actuators of Section 4.5.
; thermal-actuator6.prg
; Sawyer Fuller 4/24/00
; Builds an ink jet printed electrothermal actuator onto a glass slide.
; Manually-set origin is at edge of release layer so that contact pads are
adhered to slide.
post-print: heat at 250 C for 30 minutes before releasing to
avoid cracking
nozzle droplet frequency 125Hz
v6 = 27v9 =30v12 = 12su :initialize
v50 = v50 + 16v51 = v51 + 25su :toorigin
:heatuator
:contact-pads
; nozzle number for print head (0 is leftmost, 47 is rightmost); nozzle for photoresist (insulator) head, head 2; nozzle number for trident head (1 is rightmost, 16 is leftmost)
; add offset to where edge of relase material is
Start of heatuator section.
Origin is base of cantilever to allow easy registration
to edge of release layer. Just measure metric position on slide.
35
loop 8 ; layers of contact pads
su :toorigin ; to x,y,z coordinates specified in variables v50, vS1, and v52
v101 = 2.95 ; x and y contact pad dimension variables
v102 = .6li x=(-v10l) y=(-.6)
su :platehoriz ; build a plate structure with x-direction strokes - padi
li x3.1
su :plate-horiz ; pad2next ; layer of contact pads
:cantilevers ; cantilever section
loop 40 ; passes of cantilever structure
su :toorigin
Thick half of cantilever.
Three times as thick.
li x=(-.1) y(-.6)v104 = v232
li y-3.1
su :line-vert
y=(-2+.135) fv4
su :line-vert-backwards
y=(2-.135) fv4
su :line-vert
li y=(3.1 - 8) fv4 ; move to horiz bar position
li y-.05 ; kluge
:crossbar ; build structure at tip of cantilever
loop 3 ; 3 lines of width make up the tip
v104 = .30 ; tip made of from short line procedure
; makes the tip stiff structurally
su :shortline-horiz.here
li y-.05
next ; layer of tip structure
su :toorigin ; recenter, do the thin half of the can
li x=(.1) y(-.6)
v104 = v232su :line-vert-here
su :cure-pause ; move out of the way and pause for sol
su :toorigin
next ; cantilever layer
tilever
vent to evaporate
:donesu :endsu :allon ; turn on all nozzles once finished so they don't dry out
; v15 says which head is currently active; move new nozzle to same position over substrate; reset origin value: x; reset origin value: y; turn on dpc head to keep from drying out
:switch-to-first _dpc-head ; usually the silver ink jet head
su :alloffv15 = 0 ; which head is active
li x-v14 y=(v7 - v13) ; move new nozzle to same position over substrate
v5O = v5O - v14 ; reset origin values v50 v51 v52
v51 = v51 + v7 -v13output 3,0 ; turn on trident head to keep from drying out
return
:alloffoutput 0,1
output 2,1
output 3,1
fv4
return
:allon
output 0,0
output 2,0
output 3,0
fv1return
:cure-pausewait on
su :off
su :away
su :on
loop 1
li x1 f80
li x-1 f80
next
su :off
su :return-from-away
wait off
return
:away ; move the printhead assembly out of the way
38
if v15 = 3 :dumboffaway ; do this weird procedure if trident head is activeli x-150 y-60 fv4goto :awayend
: dumboff awayoutput 0,1
output 2,1
ii x-240 y-100 fv4
output 0,0
output 2,0
:awayend
return
:return-from-away
if v15 = 3 :dumboffreturn ; same explanation as in previous subroutine
li x150 y60 fv4goto :return-from-awayend
:dumboffreturn
output 0,1
output 2,1
li x240 y100 fv4output 0,0
output 2,0
:return-from-awayend
return
:on
fv1output v15,0
return
:off
output v15,1
fv4
return
:toorigin
su :off
velocity off
pr ab ; absolute coordinates
li zv52
li x=v50 y=v51
pr in ; back to incremental coordinates
return
:end
su :alloff
pr ab ; absolute coordinates
su :off
39
velocity off
li xO y0 zO
pr in ; back to incremental coordinates
su :allon
home x y z
su :alloff
su :toorigin
su :away
return
:shortline-horiz-here
; v104 specifies length
; couple of drops every 30 microns
; assume 125 hz drop rate
vO = 1/250*1000 ; time dwell duration in ms
v234 = cvi(v104/.060) ; integer division
loop v234
output v15,0
dwell vO
output v15,1
li x=.060
next
li x=(-v234 * .060)
return
:line.horiz
Makes a line v104 mm long starting 3.1mm away from current position.
Path ends v104+8 mm away.
f800
velocity on ; use this weird procedure to start moving gantry before
; turning on the ink nozzle so full speed is achieved before droplets
start.
li x4
output v15,0 ; on
li xv104
output v15,1 ; off
velocity off
li x4
fv4
return
:line-vert
f800
velocity on
li y4
output v15,0 ; on
li yv104
output v15,1 ; off
40
velocity off
li y4
fv4
return
:line-horiz-herePut horiz line starting from current position and ending at current position.Better for abstraction but slower.
li x-3.1 fv4
su :line-horiz
li x=(-v104 - 8 +3.1) fv4
return
:line-vert-here
li y-3.1 fv4
su :line-vert
li y=(-v104 - 8 +3.1) fv4
return
:line-horiz-backwards
f800
velocity on
li x-4
output v15,0 ; on
li x=(-v104)
output v15,1 ; off
velocity off
li x-4
fv4
return
:line-vert-backwards
f800
velocity on
li y-4
output v15,0 ; on
li y-v104
output v15,1 ; off
velocity off
li y-4
fv4
return
:trace-horiz
; trace .20 mm wide
li x-3.1
loop 2
su :line-horiz
41
li y.05 x=(-2+.135) fv4
su :line.horiz-backwards
li y.05 x=(2-.135) fv4
next
li y-.20 x3.1
return
:line-vert.here-backwards
li y3.1 fv4
su :line-vert-backwards
li y=(v104 +8 -3.1) fv4
return
:plate.horizprints a plate structure using a series of x-direction passes
v101 = width; v102 = height
v104 = v101vO = cvi(v102/.20)loop vO
su :trace-horiz
li y.20
su :cure.pause
next
li y=(-.20 * vO)
return
; integer divide
; height of plate
42
Bibliography
[1] Paul A. Williams. Three dimensional printing: A new process to fabricate pro-totypes directly from cad models. Master's thesis, Massachusetts Institute ofTechnology, 1990.
[2] David B. Wallace and Donald J. Hayes. Solder-jet technology update. SPIEInternational Symposium on Microelectronics, pages 681-684, 1997.
[3] Gregg Duthaler. Design of a drop-on-demand delivery system for molten soldermicrodrops. Master's thesis, Massachusetts Institute of Technology, 1995.
[4] Fuquan Gao. Molten Microdrop Deposition and Solidification Processes. PhDthesis, Massachusetts Institute of Technology, 1994.
[5] T. R. Hebner, C. C. Wu, D. Marcy, and et. al. Ink-jet printing of doped polymersfor organic light emitting devices. Applied Physics Letters, 72:519-521, 1998.
[6] Shun-Chi Chang, Jayesh Bharathan, Yang Yang, and et. al. Dual-color polymerlight-emitting pixels processed by hybrid inkjet printing. Applied Physics Letters,73(18):2561-2563, 1998.
[7] G6khan Perin, Thomas S. Lundgren, and Butrus T. Khuri-Yakub. Controlledink-jet printing and deposition of organic polymers and solid particles. AppliedPhysics Letters, 73(16):2375-2377, 1998.
[8] Kien F. Teng and Robert W. Vest. A microprocessor-controlled ink jet print-ing system for electronic circuits. IEEE Transactions on Industrial Electronics,35(3):407-411, 1988.
[9] J. Heinzl and C. H. Hertz. Ink-jet printing. Advances in Electronics and ElectronPhysics, 65:91-171, 1985.
[10] Brent Ridley, Babak Nivi, and Joseph Jacobson. All-inorganic field effect tran-sistors fabricated by printing. Science, 286:746-749, 1999.
[11] Brent Ridley. Inorganic semiconductors for printed transistors. Master's thesis,Massachusetts Institute of Technology, 1999.
[12] A. P. Alivisatos. Semiconductor clusters, nanocrystals, and quantum dots. Sci-ence, 271:933-937, 1996.
43
[13] Ph. Buffat and J. P. Borel. Size effect on the melting point of gold particles.Physical Review A, 13(6):2287-2298, 1976.
[14] A. N. Goldstein. The melting of silicon nanocrystals: Submicron thin-film struc-
tures derived from nanocrystal precursors. Applied Physics A, 62:33-37, 1996.
[15] H. Guckel, J. Klein, T. Christenson, and et. al. Thermo-magnetic metal flexureactuators. Technical Digest, Solid State Sensor and Actuator Workshop, HiltonHead, SC, USA, 13-16 June 1992, pages 73-75.
[16] John H. Comtois and Victor M. Bright. Applications for surface-micromachinedpolysilicon thermal actuators. Sensors and Actuators A Physical, 58:19-25, 1997.
[17] John. H. Comtois, Victor. M. Bright, and M. Phipps. Thermal micro-actuatorsfor surface-micromachining processes. Proc. SPIE, Vol. 2462, Micromachined
Devices and Components, pages 10-21, 1995.
[18] Barret Comiskey, J. D. Albert, Hidezaku Yoshizawa, and Joseph Jacobson. An
electrophoretic ink for all-printed reflective electronic displays. Nature, 394:253-255, 1998.
[19] Sawyer Fuller and Joseph Jacobson. Ink jet fabricated nanoparticle mems. In
Proceedings of IEEE Micro-electromechanical Systems, pages 138-141, 2000.