Rochester Institute of Technology
RIT Scholar Works
Theses Thesis/Dissertation Collections
8-9-1985
Characterization of G.C.A. Wafertrac 9000 using
Kodak Micro Positive 820 Resist and Developer
ZX-934
Janet Chapin McFarland
Follow this and additional works at:http://scholarworks.rit.edu/theses
This Thesis is brought to you for free and open access by the Thesis/Dissertation Collections at RIT Scholar Works. It has been accepted for inclusion in Theses by an authorized administrator of RIT Scholar Works. For more information, please [email protected].
Recommended Citation
CHARACTERIZATION OF G.C.A. HAFERTRAC 9000
USING
KODAK MICRO POSITIVE 820 RESIST AND DEVELOPER ZX-934
BY JANET C. MCFARLAND
A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE
REQUIREMENTS FOR THE DEGREE OF BACHELOR OF SCIENCE
IN THE SCHOOL OF PHOTOGRAPHIC ARTS AND SCIENCE IN
THE COLLEGE OF GRAPHIC ARTS AND PHOTOGRAPHY OF THE
ROCHESTER INSTITUTE OF TECHNOLOGY.
Janet McFarland
SIGNATURE OF AUTHOR: ---:=~::-::-=-~=--=~==-::==-=-=:-:-:=-=--=======
_ _
_
IMAGING AND PHOTOGRAPHIC SCIENCE
CERTIFIED BY: _ _ _ _ _
L...:y~n=_=n=_=F:_::_u__::_::;II
,:,;,e=::r
~
_ _ _ _
_
THESIS ADVISORTHESIS RELEASE PERMISSION FORM
ROCHEsTER INSTITUTE OF TECHNOLOGY
COLLEGE OF GRAPHIC ARTS AND PHOTOGRAPHY
TITLE OF THESIS: CHARACTERIZATION OF A G.C.A. WAFERTRAC 9000
USING KODAK MICRO POSITIVE 820 RESIST AND DEVELOPER ZX-934
I,
Janet McFarland
, hereby grant permission to the
Wallace Memorial Library of Rochester Institute of Technology
to reproduce my thesis in whole or in part. Any reproduction
will not be for commercial use or profit.
Janet McFarland
SIGNATImE: ________________________________________ __
DA~:
___________~8~,V~
·
y
~
-
/~
8
~
~
L
-
CHARACTERIZATION OF A G.C.A. WAFERTRAC 9000
USING
KODAK MICRO POSITIVE 820 RESIST AND DEVELOPER ZX-934
By Janet C. McFarland
ABSTRACT
A static coating program and a spray/ static development
program were developed, for use on the Wafertrac 9000
using Kodak Micro Positive 820 resist and developer ZX-934 in order to minimize critical dimension shift
and maximize process latitude. By maximizing process
latitude, a greater tolerance for process and equipment
errors can be achieved. Using the established process,
1.2um resist lines were produced and critical dimension
shift was minimized providing a process with wide
process latitude for creating photoresist geometries
with dimensions larger than 1.2um.
ACKNOWLEDGEMENTS
The author would like to thank those people who
contributed to the successful completion of this thesis.
A thank you to Dr. Lynn Fuller, the Director of the
Microelectronic Engineering Program, for supervising my thesis.
A special thanks must be extended to Scott Blondell,
for without his expertise in maintaining the equipment and
in rapid repairs, there would have been no feasible way this thesis would have been completed.
Thank you to the Microelectronic Engineering Program,
for permitting me to work and use the equipment in the labs.
Finally, thank you to all those people who listened to
me complain, made me smile when all looked bleak, and
supported me through my work as an Imaging Scientist. Without those people life would have been quite unbearable.
TABLE OF CONTENTS
I. INTRODUCTION 1
II. EXPERIMENTAL 9
A. Equipment and Material Procurement 9
B. Preparing Wafers for Resist Coating 9
C. Preparing Line 3 for Resist Coating 10
D. Resist Coating 10
E. Resist Coating Evaluation 11
F. Exposure 12
G. Preparing Line 2 for Developing the Wafer 14
H. Processing Wafer 15
I. Evaluating Linewidths 15
III. RESULTS 17
A. Resist Thickness vs Spin Speed 17
B. Critical Dimension Shift vs Exposure Energy . 18
C. Critical Dimension Shift vs Position on Wafer 20
IV. DISCUSSION 21
V. CONCLUSION 23
VI. REFERENCES 24
APPENDIX A: Photolithographic Process
APPENDIX B: Control Programs for the Wafertrac 9000
APPENDIX C: Resist Thickness Uniformity vs RPM
APPENDIX D: Exposure/Development Matrix
VITA
LIST OF FIGURES
FIGURE 1: Top veiw of the G.C.A. Wafertrac 9000 3
FIGURE 2: Interferogram. Plot of Reflected Beam
Intensity versus Time 7
FIGURE 3: Plot of Oxide Thickness versus Time for
a Steam Environment 9
FIGURE 4: The design of evaluating resist thickness
using the principle of single beam
interf erornetry 12
FIGURE 5: Relative Output of a High Pressure
Mercury Vapor Lamp 13
FIGURE 6: R.I.T Resolution Target manufactured by
Gould/A. M.I 14
FIGURE 7: Cross sectional view of a measured
linewidth 15
FIGURE 8: Resist Thickness vs Spin Speed 17
FIGURE 9: Critical Dimension Shift vs Exposure
Energy (development time 30 sec.) 18
FIGURE 10: Critical Dimension Shift vs Exposure
Energy (development time 35 sec.) 18
FIGURE 11: Critical Dimension Shift vs Exposure
Energy (development time 40 sec.) 19
FIGURE 12: Critical Dimension Shift vs Exposure
Energy (development time 50 sec.) 19
FIGURE 13: Critical Dimension Shift vs Position
on Wafer (2um line measured) 20
FIGURE 14: Critical Dimension Shift vs Position
on Wafer (5um line measured) 20
LIST OF TABLES
TABLE 1: Observed Colors for different Thin Film
Thickness for Perpendicular Illumination
with White Light 11
I. Introduction
For the past three decades, industry has been
continually investigating means to miniaturize electronic
devices. In 1965, only 250 devices could be interconnected
in the "monolithic integrated circuit"; whereas, in 1983
over 1,000,000 devices could be interconnected in a single
integrated circuit.C 13 This improvement is primarly due to
the decrease in size of circuit devices. This can be
illustrated by the decrease of typical linewidth from 20um in 1963 to 2.0um in 1983. CI 1 These advances in
miniaturization were achieved by improvements in the fabrication techniques and equipment.
The desire to achieve the smallest device for the
cheapest price has continually driven the research and
development departments of companies to uncover new ideas.
In 1979, the Research and Development Department of GCA/IC
System Group was founded on the idea of incorporating the
photolithographic process into the automated,
computer-controlled Wafertrac system.C23
The photolithographic process is a process by which
silicon wafers are treated in order to build integrated
circuits. The photolithographic process begins with an
oxide, which is initially grown on the wafer's surface. This thermal oxide provides a stable, uniform protective
layer on the silicon wafer.C33 The oxide layer also creates
a hydrophilic surface on the wafer. Such a surface is
is coated with photoresist and softbaked. Softbaking is
done in order to stabilize the solvent content and to obtain
reproducible resist thickness. C 33 The wafer is exposed, in
order to produce a latent image of a desired pattern. The
wafer is then developed and postbaked to increase the
resistance of the resist to the etching process.C 33 Finally,
the oxide is etched and the photoresist is stripped.
Etching is done to remove the oxide layer left exposed in
the development process. A pictorial example of the overall
processing can be found in Appendix A.
The Wafertrac system incorporates the photolithographic
process and offers reliable and repeatable results. With
proper programming, resist uniformities have been repeatedly
demonstrated as being 50 angstroms across the wafer
surface, along with dispensing photoresist within +2%
repeatability over a range of 2.0 - 9.0
cc.C53 All of this
is quite difficult for a human operator to duplicate by hand
spinning of resist and developer. Furthermore, without
precision and reproducible results, it would be difficult to
reproduce repeatedly small line geometries.
The Wafertrac 9000 is a system which was specifically
designed for the Microelectronic Engineering Labs at
Rochester Institute of Technology. The system consists of a
series of buffers; a high pressure scruber, which exerts
approximately 2400 psi of deionized water; a photoresist
coating chuck; a series of hot plates; along with a positive
and a negative developer spin chuck. The entire system is
convection oven
-^
FIGURE li Top view of the G.C.A. Wafertrac 9000
Previously, no work had been done to characterize the
Wafertrac 9000 system at R.I.T. The resist, Kodak Micro Positive 820, and developer ,ZX-934, were designated as the
chemical agents for the system. The Kodak Micro Positive
820 resist is diluted at a rate of lOOcc of resist to 33cc
of Kodak Micro Positive Thinner. This dilution gives a 32%
solids content. Kodak Micro Positive Developer is diluted
1:1 with deionized water. The purpose of looking at the system and the process simultaneously is to be able to explain the system's parameters more efficiently along with
establishing a process which will produce repeatable lines
below 2um in the Microelectronic Engineering labs at Rochester Institute of Technology.
Initially, linewidth is established by the mask which is used, but it is possible to alter linewidth dimensions by changing various photoresist process parameters.C63 The
energy, and development time. The postbake, etching and
stripping processes of the photolithographic process were
not examined within the scope of this work.
The technique by which photoresist is applied is very
important. Photoresist thickness and uniformity are
products of the speed of the spinner. Low spin speeds have
a tendency to build up resist along the edge of the wafer,
whereas high spin speeds tend to spread the resist
unevenly.C73 It is recommended, therefore, that spin speeds
be held between 3000 RPM and 6000 RPM.C83
Uniformity across the wafer's surface is a function of
how the resist is applied and the condition of the surface
of the wafer. Two methods of dispensing resist are static
dispense and spin coating. The static dispense technique
applies resist in a puddle, which covers 75% of the wafer's
surface, then accelerates to a spin speed in order to spread
the resist over the remaining surface. In contrast, the
spin coating method dispenses resist as the wafer Is
rotating. By carefully balancing the dispense time of
resist, the acceleration and the spin speed of the wafer,
one can achieve a uniform resist coating with either
process.
Uniformity across the wafer is very important in
linewidth dimensions. The result of an uneven surface is
either overexposure or underexposure of the resist. This
has a lasting effect on the remaining steps in the process.
The effects can be seen in the change on linewidth
The Wafertrac coating modules can be programmed to
accommodate either of the programs. The module has the
capability of accelerating and decelerating at a maximum
speed of 50,000 RPM/sec for a 76 mm wafer. Time intervals
are programmable from 0.1 seconds to 25.4 seconds,
incremented by 0.1 seconds. Longer time intervals can be
obtained through multiple step programming.C 93
Spin speed is not the only parameter that affects
resist thickness. Softbake also plays a key role in
determining resist thickness. This was proven by Shaw and
Dill, who showed that two photoresist samples of equal
thickness before softbake did not necessarily have the same
thickness after softbaking if time or temperature was
altered.C103
Softbake is needed in order to stablize the solvent
content and to obtain reproducible resist thickness.CI 13 A
softbake at a low temperature and time would therefore not
reduce the solvent count to an acceptable limit. This may
imply that the higher and longer the softbake, the better it
removes solvents and promotes adhesion. This is partially
true, but there are trade-offs. One in particular is that
as softbake time and/ or temperature increases,
photosensitivity decreases; consequently, exposure and
development time must be increased to obtain clearing.C63
This is due to the fact that the surface of the positive
resist Is dehydrated by too lengthy a time and too high a
temperature. Temperatures exceeding 100 Celsius cause
It is therefore recommended that softbake times be held
under 100 seconds and temperature below 100 degrees
Celsius.C43
The hot plates on the Wafertrac have vacuum ports to
clamp the wafers firmly. This promotes the transfer of heat
and reduces interface resistance.
These processes and parameters are contained and
controlled on Line 3 of the Wafertrac 9000. Line 3 also
contains a high-pressure scrubber and a dehydration oven,
along with the resist coating module and hot plate for
softbaking. The dehydration bake removes water on the
surface of the wafer, thereby promoting adhesion of the
photoresist.
In order to evaluate resist thickness at the R.I.T.
labs, a technique was developed to measure resist thickness
by single beam interferornetry. Single Beam interferometry
works on the principle that when a monochromatic light is
reflected off a surface coated with a thin film, the
reflections from the two surfaces will interfere, producing
a reflected beam with an intensity that depends on the
thickness of the thin film. The optical path for the beam
reflected from the silicon surface is 2dn/N longer than the
beam reflected from the surface of the thin film. The
resist coating serves as the thin film. Construction
interference occurs when this path difference is equal to NX
thus:
NX = 2dn where N = 1,
gives the condition for peaks in the reflected beam
intensity. By rearranging the equation one obtains the
equation for determining the thin film thickness;
d = NX/2n
where d: thickness of the thin film
n: index of refraction for the thin film
X: wavelength of the monochromatic light
N: number of cycles
The number of cycles is determined by an interferogram (plot
of reflected beam intensity versus time) an example of is
shown below.
intensity Thickness=
)> N
2 n
time
N : number of cycles
n : 1.622
FIGURE 2: Interf erogram. Plot of Reflected Beam
Intensity versus Time.
It is by this technique that the thickness of the resist can
be determined.C123
It is well known that the exposure will affect
linewidth change. The exposure and development process are
interrelated. One can optimize a development process for
8
against linewidth change.C33
The spin chuck for development has the capability of
dispensing deionized water and positive developer. The
deionized water is used in order to wet the surface of the
wafer so that uniform development will occur over the wafer's surface.C133
Through the process of optimizing the development and
exposure steps for a prescribed resist thickness, one can achieve minimal critical dimension shift and wide process
latitude. This was a desired result in characterizing the
II. EXPERIMENTAL
A. Equipment and Material Procurement
The majority of the equipment, materials, and lab
facilities were provided by the Microelectronic Engineering
Labs at Rochester Institute of Technology. These include
the Wafertrac 9000, Kasper 2000 Mask Aliner, Metrologic
Radiometer,Tegal Plasmaline 200, Metrologic Helium Neon
Laser, International Light IL-1500 photoresist radiometer,
EAI 1125 Variplotter, RIT Resolution Target, and Nanometrics
Nanoline III. In addition, Kodak Micro Positive 820 Resist,
Thinner, Developer ZX-934, and silicon wafers were provided
by the Microelectronic Engineering Labs.
B. Preparing Wafers for Resist Coating
rcr1
1 1 I II I I II 1 1 I I I I 1 1 1 1 1 I II I I
I ll IIII | | | | |
to 100
I, tim.(mini
FIGURE 3: Plot of Oxide Thickness verses
10
Before the resist was applied on the wafers a 1200
angstrom layer of silicon dioxide was grown on the wafers.
The oxide was grown in a steam environment at 1100 degrees
Celsius for 30 minutes. The average oxide thickness was
then evaluated with an elipsometer at 1240.50 Angstroms.
C. Preparing Line 3 for Resist Coating
Line 3 of the Wafertrac 9000 consists of 4 modules:
the high pressure scrub, convection oven, coating module,
and hot plate. The modules were programmed. Individual
programs can be found in Appendix B.
The high pressure scrub was programmed to exert 2400
psi of deionzed water. The convection oven, which serves as a dehydration bake after the high pressure scrubber, has three temperature zones, each programmed at 300 degrees
Celsius. The wafer travels through each zone at 0.37
mm/ sec. The base program used for dispensing resist was a
static dispense program. The resist was dispensed for 1.2
seconds, then accelerated to 500 RPM, and finally spun at a
terminal speed. Terminal speeds were later varied to determine resist thickness versus spin speed. Finally, the hot plate was programmed at 95 degrees Celsius for 60
seconds. This final step on line 3 serves as the soft bake
step.
D. Resist Coating
Wafers were processed on Line 3. A resist thickness versus
11
spin speed from 3500 RPM to 6500 RPM. A speed of 5000 RPM
produced a 1.2 um resist thickness. This was the selected
resist thickness for the remaining experiment.
E. Resist Coating Evaluation
The uniformity of the resist was evaluated visually. A
change of color indicates a change in the resist thickness.
For example, a color change from blue to orange across the
wafers surface would indicate a quarter of a wavelength
change in thickness across the wafers surface.
COLOR
(A color change represents 1/12 of a
wavelength difference)
Violet
Blue
Green
Yellow
Orange
Red
Violet
TABLE 1: Observed Colors for different Thin Film Thickness
for Perpendicular Illumination with White Light.
12
Resist thickness was evaluated using the principle of
single beam interferornetry. A wafer was placed in a plasma
asher; a Helium-Neon laser beam was reflected off the wafer'
s surface onto a detector which was connected to a x-y
plotter, as shown in Figure 4.
As the wafer was ashed, the reflected intensity versus
time was recorded on the x-y plotter. The interferagram was
then evaluated and the thickness determined as indicated in the introduction.
it* A.
sector
Wafer
FIGURE 4: The design of evaluating resist thickness using
the principle of single beam interf erornetry
-F. Exposure
A Kasper Maskaliner equiped with a High Pressure
Mercury Vapor Lamp was used to expose the wafers. An
exposure range of 45 mJ/cm^to 85
mJ/cm*
13
give a wide range of exposure effects. This range was then
used in ascertaining the proper exposure needed for a
specific technique of development.
The light source used was a high pressure mercury vapor
lamp. The relative output is shown in figure 5 with the
resist absorption of the Kodak Micro Postive 820 resist
superimposed over it. The radiometer which was used to
measure irradiance at the wafer's surface had a simular
responce to the resist absorption curve. A black glass was
also used when measuring irradiance in order to simulate the
same conditions when patterning the wafer with the
resolution mask
Abaorptton
<,U.Sr.,M..v*n y MO
11
MO MO
Wnnflth(mn)
TOO
FIGURE 5: Relative Output of a High Pressure
Mercury Vapor LampC153
An R.I.T. resolution was used to pattern the wafers.
14
designed to range from 0.5um to 50.0um.
$2= -)!i-*J'_P _ "=lilii-=IiIiIiI^i,ui,i:
uiiitluditwiiwi^iriji:
FIGURE 6: R.I.T. Resolution Target manufactured
by Gould/A. M.I.
G. Preparing Line 2 for Developing the Wafers
Line 2 consists of 2 modules, the spin chuck for
development and the hot plate for postbaking. A base
program for development was established. The program was a
spray/static development routine, which means the developer
was sprayed onto the wafer's surface, while the wafer was
spinning. After a predetermined time elapsed the wafer
stopped spinning and a meniscus was formed on the wafer's
surface, the developer then remained for another set time.
The initial spray development time was programmed for 4
seconds. The final static development time was varied with
exposure in order to generate an exposure/development
15
H. Wafer Processing
The wafers were then processed through Line 3 and Line
2 using the previously described programs. In order to
generate the exposure/development matrix, exposure energy
was varied from 45-85 mJ/cm and development time was varied
from 30-50 seconds. For each step in the matrix five wafers
were processed for total of 80 wafers.
I. Evaluating Linewidths
Linewidths were evaluated on a Nanometrics Nanoline III
linewidth measurment system. The Nanoline was calibrated to
use the 5.0um and 2.0um line on the mask as the standard.
Linewidths were measured from minimum to minimum, as shown
in Figure 7.
f
Ajj^jj
J
r
\
K W^-width
FIGURE 7: Cross section view of a measured
linewidth.
16
repeated 3 times. These values were then averaged.
Critical dimension shift was determined by subtracting the
average photoresist linewidth from the 5um and 2um values
17
III. RESULTS
A. Resist Thickness vs Spin Speed
L500
g L350
3,
cn tn \u z
O 1200
z
1.050
RESIST THICKNESS vs SPIN SPEED
KODAK Micro Positive Resist 820, Viscosity32%
+- -1- -+-
-+-3000 3500 4000 4500 5000 5500 6000 6500 7000
SPIN SPEED (RPM)
18
B. Critical Dimension Shift vs Exposure Energy (with varying development time)
CRITICAL DIMENSION SHIFTv EXPOSURE
ENERGY KOOAK Micro PositiveResist 820 KODAK Mico FbsitiveDeveloper ZX-934
0.70+ E 3 5 o.oo -0.70 < o 1.40
40.00 50.00 60.00 70.00 EXPOSURE ENERGY (mj/cffl2)
AverageResist Thickness Belore Development: 1.2um
DevelopmentCone.1 dev:1water
DevelopmentTime: 30see
30.00 90.00
FIGURE 9: Critical Dimension Shift vs Exposure Energy (development time 30 sec)
CRITICAL DIMENSION SHIFT EXPOSURE ENERGY KODAK Micro Positive Resist 820 KODAK Mico PositiveDeveloper ZX-934
- 0.50 5 0.00-u S 5 < u S u 0.50 -1.00 40.00
AverageResist Thickness Belore Development: 1.2um Development Cone.1 dev: 1water DevelopmentTime:35sec
50.00 60.00 70.00 EXPOSURE ENERGY (mj/cm2)
30.00 90.00
CRITICALDIMENSION SHIFT vs EXPOSURE ENERGY
KODAK MicroPositive Resist 820 KODAK MicoPositiveDeveloper ZX-934
19 f -55 t E o 0.00--0.55 -1.10+
AverageResist ThicknessBefore Development: 12um DevelopmentCone.1 dev:1water
DevelopmentTime: 40ec
!0.00 50.00 60.00 70.00 80.00
EXPOSURE ENERGY (mj/cm2)
90.00
FIGURE 11: Critical Dimension Shift vs Exposure Energy
(develpment time 40 sec)
CRITICAL DIMENSION SHIFT vs EXPOSURE ENERGY
KODAK MicroPositiveResist 820 KODAK MicoFbsitiveDeveloper ZX-934
_ 0.90+ E 0.00-z 5 -0.90 E u -i.eo
40.00 50.00 60.00 70.00 EXPOSURE ENERGY (mJ/on2)
Average Resist Thickness Before
Development: i.2um DevelopmentCone.1dev: 1 water
DevelopmentTime: SOsec
30.00 90.00
FIGURE 12: Critical Dimension Shift vs Exposure Energy
20
C. Critical Dimension Shift vs Position on Wafer
0.30+
CRITICAL DIMENSION SHIFT "
POSITION ON WAFER KODAK Micro Positive Resist 820
KODAK Mico PositiveDeveloper ZX-934 AverageRacistThicknessBetora
Development:1.2um Development Cone.1 dev: 1water
DevelopmentTime: 40sec
Linewidth Measured:2um
10.00 20.00 30.00 40.00
POSITION ON WAFER (mm)
50.00 60.00
FIGURE 13: Critical Dimension Shift vs Position on the
Wafer (2um line measured)
_ 0.3C z o 55 z in X 5 _i < u S u 0.1! 0.00 -0.15-10.00
CRITICAL DIMENSION SHIFTvs POSITION ON WAFER
KODAK Micro Positive Resist 820 KODAK Mico FbsitiveDeveloper ZX-934
Average Resist Thickness Before
Development:1.2um DevelopmentCone.1dev:1water
DevelopmentTime: 40sec Linewidth Measured: Sum
20.00 30.00 40.00 50.00 60.00
POSITION ON WAFER (mm)
FIGURE 14: Critical Dimension Shift vs Position on the
21
IV. Discussion
A static dispense coating program and a spray/static
development program were refined as a result of the
completed research. These programs were designed in order
to establish wide process latitude and minimal critical
dimension shift.
The static dispense coating program, dispenses resist
for 1.2 seconds then accelerates to 5000 RPM at a rate of
40,000 RPM/sec, spinning for 10 seconds. The final spin
speed was determined by examining figure 4 and Kodak graph
of Resist Thickness Uniformity vs RPM, found in Appendix C.
To produce a thickness of 1.2um a final spin speed of 5000
RPM was selected.
The spray/ static development program dispenses
deionized water in order to wet the surface of the wafer,
then spins at a speed of 500 RPM while developer is sprayed
on the wafer's surface for four seconds. Spinning is then
terminated and a meniscus is formed on the surface of the
wafer. The meniscus of developer remains for 36 seconds and
then the wafer is rinsed with deionized water.
The total development time of 40 seconds was choosen
since the slope of the line changed the least around the
zero critical dimension shift, as seen in figure 11. Thus
some process latitude is gained since the exposure can range
from 50 mJ/cnrHo 60 mJ/cm*without a significant difference
in critical dimension shift, compared to the 50 second
22
Thus, in order to have a critical dimension shift of zero
there would be only one value for exposure.
Figures 13 and 14 represent the critical dimension
shift across the wafer. Notice in figure 13 that the
critical dimension shift is not centered around zero. This
effect was due to the fall off of irradiance towards the
edge of the illuminated area. The irradance ranged from 4.5
to 5.0 mW/cm*
where 5.0 mW/cm2
was at the center of the
illuminated area and 4.5 mW/cm^was found at the edge. 5.0
mW/cnr: The effect is less noticable in figure 14, due to
the fact the linewidth measured was larger than that in
23
V. CONCLUSION
The process which was established on the Wafertrac
9000, using Kodak Micro Positive 820 resist and developer
ZX-934, produced minimal critical dimension shift and gave
some process latitude in the exposure step of the process.
Resist lines of 1.2um were also produced using this process.
Recommendations for future work are as follows:
1. Examine other methods for development, such
as a spray/ static/spray and spray rountine.
2. Optimize softbake time and temperature for
the over all process.
3. Complete optimizing the photolithographic
process for wide process latitude, by examining
the postbake, etching and stripping effects on
the process latitude and critical dimension
24
IV. References
1. L.F. Thompson, C.G. Willson, and M.J. Bowden,
INTRODUCTION TO MICROLITHOGRAPHY, American
Chemical Society, Washington, D.C, 1983 (pp.IX).
2. J. A. Irwin and T.J. Weber, PROCEEDINGS OF KODAK
INTERFACE '81 MICROELECTRONICS SEMINAR, G-135, (1981).
3. D.J. Elliott, INTEGRATED CIRCUIT FABRICATION TECHNOLOGY,
McGraw-Hill Co., New York, N.Y., 1982.
4. L.F. Fuller, Private Communication, Rochester Institute
Technology (Oct. 1984).
5. GCA Corporation, SOLID STATE TECHNOLOGY, 26, 69 (1983).
6. E.B. Hryhorenko, PROCEEDINGS OF KODAK INTERFACE '80 MICROELECTRONICS SEMINAR, G-130, (1980).
7. M.W. Chan, PROCEEDINGS OF KODAK INTERFACE '75 MICROELECTRONICS SEMINAR, G-45, (1975).
8. P. O'Hagan and W.J. Daughton, PROCEEDINGS OF KODAK INTERFACE '77 MICRELECTRONICS SEMINAR, G-48, (1977).
9. WAFERTRAC APPLICATION NOTE, 4/82 4M CPC (1982 GCA).
10. F.H. Dill and J.M. Shaw, IBM JOURNAL RES. DEVELOPMENT.,
21, 210 (1977).
11. T. Batchelder and J. Piatt, SOLID STATE TECHNOLOGY,
26, 211 (1983).
12. L.F. Fuller, Private Communication, Rochester Institute
Technology (Jan. 1985).
13. WAFERTRAC APPLICATION NOTE ADDENDUM, 10/82 6M CPC
(1982
GCA)-14. R.A. Colclaser, MICRO ELECTRONICS PROCESSING AND DEVICE
DESIGN, John Wiley & Sons, New York, 1980.
15. V. Miller and H.L. Stover, SOLID STATE TECHNOLOGY,
APPENDIX A: PHOTOLITHOGRAPHIC PROCESS
silicon
The Wafer
Oxide Growth
2 Si02
B silicon
_ photoresist
I SiO,
Photoresist Coat
and Soft Bake
silicon
light
Exposure
..;r:'y.i-.\
.
...i
photoresist
|.- ::
Si0
silicon
After Development
5
Etching Oxide
photoresist
Si02
silicon
Si02
silicon
APPENDIX B: Control Programs for the Wafertrac 9000
STATIC DEPENCE COATING PROGRAM
STEP NO. FUNCTION VALUE
10 11 20 SEC/10
20 6 1 ON(photoresist)
30 11 12 SEC/10
40 6 0 0FF(photoresist)
50 11 25 SEC/10
60 9 200 RPM/sec
70 10 400 RPM
80 11 20 SEC/ 10
90 9 40000 RPM/sec
100 10 5000 RPM
110 11 100 SEC/ 10
120 13 0 LATCH
130 11 100 SEC/10
140 10 0 RPM
150 16 5050 RPM
160 17 5150 RPM
FTBAKE PROGRAM
STEP NO. FUNCTION VALUE
10 1 95 zone 1; 95 C
20 11 60 SEC
DEVELOPMENT PROGRAM; SPRAY/ STATIC
STEP NO. FUNCTION VALUE
10 2 1 ON
20 11 50 SEC/10
30 2 0 OFF
40 9 500 RPM/ SEC
50 10 500 RPM
60 11 20 SEC/10
70 3 1 ON
80 10 500 RPM
90 11 40 SEC/10
100 10 0 RPM
110 11 20 SEC/10
120 3 0 OFF
130 11 360 SEC/10
140 2 1 ON
150 10 500 RPM
160 11 50 SEC/10
170 10 500 RPM
180 11 50 SEC/10
190 2 0 OFF
200 10 5500 RPM
210 11 100 SEC/10
HIGH PRESURE SCRUB PROGRAM
STEP NO. FUNCTION VALUE
10 9 10000 RPM/SEC
20 10 3000 RPM
30 1 1 ON
40 8 1 ON
50 11 250 SEC/10
60 1 0 OFF
70 8 0 OFF
80 10 4000 RPM
90 2 1 ON
100 11 50 SEC/10
110 2 0 OFF
120 10 0 RPM
130 11 20 SEC/10
CONVECTION OVEN PROGRAM
STEP NO. FUNCTION VALUE
10 1 300 zone 1; 300 C 20 2 300 zone 2; 300 C
30 3 300 zone 3; 300 C
APPENDIX C: RESIST THICKNESS UNIFORMITY vs RPM
RESISTTHICKNESS UNIFORMITY vsRPM
APPENDIX D: EXPOSURE/DEVELOPMENT MATRIX
50 AAA
40 * A A A
DEVELOPMENT
TIME 35 a a a
(SEC)
30 A A A
45.0 60.0 70.0 85.0
EXPOSURE ENERGY (mJ/cm )
VITA
Janet Chapin McFarland was born in New Castle, PA. She
was raised in Wexford, PA, a suburb of Pittsburgh, and
attended North Allegheny High School. After spending a year
abroad studying in Zug, Switzerland as a A.F.S student, she attended Rochester Institute of Technology to pursue a
Bachelor of Science degree in Imaging and Photographic
