Beam Angle Control on the VIISta 80 Ion Implanter
Christopher Campbell, James Cummings, Robert Lindberg, Joseph C. Olson,
Svetlana B. Radovanov, Donna L. Smatlak
Varian Semiconductor Equipment Associates, Inc.
Gloucester, MA 01930 USA
Abstract—Advanced integrated circuit design requires precise
control of beam incidence angle. This requirement has led to the
development of an automated angle control system on Varian
Semiconductor’s high current VIISta 80 ion implanter.
In this paper we show beam incidence angle and angular spread
measurements for 200 and 300 mm ion beams on the VIISta 80
ion implanter. Multiple beam measurements are sampled across
the wafer plane for each beam setup. Beam angle computation
results are compensated for prior to wafer implantation for
optimal incident angle control. Beam, bare wafer and device
performance data were used to confirm the accuracy of this
measurement and control system. Excellent measurement
accuracy and repeatability has been demonstrated.
Data will be shown which includes arsenic, boron and phosphorus
implants from both drift and decel operation. Benefits and
process differences will be shown with active beam angle
correction as compared to classical open loop methods.
Mechanical tilt angle accuracy, repeatability and verification data
will also be discussed.
I. INTRODUCTION
Implant & Dose
Control
- DoseMap
-Tilt/Twist Positioning
-Scan Control
- Angle Correction
-Implant Process Monitoring
Implant System
- X-Tilt Subsytem
- Y-Tilt Subsystem
- Scan Subsystem
Beam Measurement
-Beam Profile
-Beam Angles
-Dosimetry
Fig. 1. VIISta 80 implant system design.
CMOS device technology scaling has imposed stringent
beam incident angle and parallelism requirements on modern
high current implanters. Classical process matching methods
utilizing sheet resistance, Thermawave and SIMS have proved
to be unreliable for matching the device parameters for modern
CMOS technology nodes [1]. Three-dimensional CMOS
effects that are sensitive to ion beam angle characteristics
during the implant process mandate angle precision that must
be measured and controlled. The VIISta 80 high current
incident angle control system has been developed to address
+45 Degree
Xtilt Angle
Process
Chamber Wall
+
Angle
Directi
on
-
+
Bea
-
Y-Tilt Angle
-
+
+
Bea
0 Degree Tilt
Fig. 2a Y-Tilt
a) Control
Fig. 2bX-Tilt
Control
b)
Fig. 2. VIISta 80 tilt control – a) Y-tilt (top view) b) X-tilt (side view).
these current and future technology node requirements. In this
paper we discuss the beam angle control system hardware and
algorithm, and present results of mechanical angle
measurement and post-implant wafer characterization.
II. BEAM ANGLE CONTROL SYSTEM
The VIISta 80 automatic beam angle control system design
is shown schematically in Fig. 1. This system incorporates
state-of-the art closed loop measurement, control and process
critical monitoring capabilities. Ion beam incident angle and
parallelism is measured and dynamically compensated for
within the implant control system. The x and y tilt are
precisely controlled, monitored and verified with integrated
traveling
profile Faraday
Control
accuracy
Y-tilt control
< ± 0.1°
Faradays cups
Beam angle
measurement
/ control
Fig. 3. Profile Faraday shadowing of sample dose Faradays
590
580
570
560
550
Fig. 4. B+, 5keV sample Faraday dip profiles from profiler
shadowing
540
530
and in-situ position sensing.
520
The VIISta 80 angle control performance is a result of a
multistep process. Initially, precision laser calibration of the
x-tilt, y-tilt wafer positioning sub-systems and beam profiler is
performed to establish an accurate reference system. This is
done as part of initial tool setup and preventative maintenance.
Next, for each implant recipe setup, beam angle
measurements are collected after the beam has been tuned for
optimum uniformity. This is accomplished by analyzing data
collected from an array of Faraday cups positioned behind the
wafer plane (Fig. 3). Beam current as a function of profiler
position is gathered as the beam profiler shadows each sample
cup Faraday. Five sampling cups are used for 200 mm
implants and 7 sampling cups are used for 300 mm implants.
Shadowing creates a dip profile for each of the sample Faraday
cups.
The displacement of each dip profile (Fig. 4) is used to
compute the beam angle at discrete positions across the wafer
[2]. The beam angle measurements are averaged (beam angle
mean) and the largest angle deviation from the mean is
computed (beam angle spread). The beam angle spread is
interlocked to a value set in the recipe by the customer. The
beam angle mean is compensated during implant using a
corresponding value of y-tilt. This compensation ensures that
the mean incident angle of the beam is normal to the wafer.
III. EXPERIMENTAL MEASUREMENTS
X and y tilt angle accuracy and repeatability data were
collected with a laser auto-collimator to confirm the
mechanical precision of the system. This 2-axis laser system
was mounted on the process chamber for direct line-of-sight
measurements of the wafer surface. The x-tilt and y-tilt
mechanical positioning systems were sequenced through
510
-2
-1
0
2
1
3
4
5
Y Tilt (deg)
Fig. 6 . B+ 60 keV 1E14 Y-tilt Channel test results without
angle correction
several discrete angle positions for 20,000 cycles while laser
angle measurements were collected. The repeatability of the
x-tilt angle was determined to be ±0.03° and the y-tilt angle
repeatability was measured to be ±0.05°.
The dip profiles for a 300 mm B+, 5 keV 11 mA recipe
are shown in Fig. 4. The data from this recipe indicate a beam
angle mean of +3.71° and a beam angle spread of ±1.15°. Fig.
5 shows beam angle data as a function of sampling cup
position with and without angle mean compensation. Note that
prior to angle compensation, angles as large as +4.77° are
measured relative to wafer normal. Angle compensation
eliminates the mean angle error and minimizes the overall
angle variation across the wafer to ±1.15°.
Beam angle integrity is measured by performing a series of
implants using a channeling sensitive recipe (in this case B+ 60
keV 1e14) while setting the y-tilt offsets for each wafer (in this
case from 4° to + 4°). The maximum channeling occurs when
the sheet resistance is at its minimum, thus indicating the
optimum beam/wafer incidence angle. Desired angle
performance is indicated by a repeatable value of this
minimum at or near 0°.
Illustrated in Fig. 6 are the y-tilt channel results collected
across multiple B+ 60 keV setups without closed loop angle
590
580
570
Beamlet Angle (Degrees)
6
5
4
3
2
1
0
-1
-2
560
w/o Angle Control
550
w/ Angle Control
540
Wafer Plane
530
520
Cup 1
Cup 2
Cup 3
Cup 4
Cup 5
Cup 6
Cup 7
Angle Cup Position
Fig. 5. B+ 5 keV 11 mA beam angle measurements
510
-2.5
-2
-1.5
-1
-0.5
0
0.5
1
1.5
2
2.5
Y Tilt (deg)
Fig. 7 . B+ 60 keV 1E14 Y-Tilt channel test results with
angle correction
505
Sheet Rho (Ohms/Sq.
1000
950
900
485
475
5 Months
465
Wfr to Wfr.
Repeatability
= 0.49
Wfr to Wfr. Repeatability = 1.11
455
850
Angle Control
No Angle Control
495
750
Rs
2
1
0
-1
X-Tilt Angle
-2
Fig 8. B+ 70 keV 5E13 X-tilt channel test results with angle
correction
control. Large differences in incident angle (>2°) are
exhibited which result from typical variations in beam tuning.
The process benefits of the VIISta 80 angle measurement
and control system are clearly evident in Fig. 7. Y-tilt channel
results from the identical implant test matrix of Fig. 6 show
that incident angle offsets are nulled out and incident angle
variations are improved by a factor of two. Fig. 8 shows x-tilt
channel results indicating negligible x-tilt beam incident angle
offset and variation. These process control performance
benefits have been demonstrated across diverse recipe sets,
recipe energies (5-70 keV), beam currents and wafer sizes over
several months.
The sheet resistance maps shown in Fig. 9 demonstrate the
on wafer process advantages when compared to wafer maps
from a corresponding non-angle corrected -2° implant. Higher
average sheet resistance indicates less channeling due to -2°
beam incident angle error. Differential channeling patterns
resulting in degraded wafer uniformity are shown, further
indicators of large beam angle variation.
Performance advantages have also been demonstrated for a
customer process monitoring recipe (P+ 70 keV 1e14, 0°/0°).
Fig. 10 shows long term dose repeatability improvements of
#
LCL RunUCL
69
65
61
57
53
49
45
41
37
33
29
25
21
17
9
1
13
445
800
5
T-Wave Mean (Tw Units)
1050
TARGET
Fig. 10. Dose repeatability improvement of >2X shown from
customer process monitor recipe when using VIISta 80 with angle
control.
greater than a factor of two for this angle sensitive process
control recipe.
SIMS profiles of a B+, 70 keV 3e13 0°/0° implant (Fig. 11)
from points taken at the top, bottom, left, center and right
wafer locations from a VIISta 80 with incident angle control
demonstrates the parallel ion beam angular precision. All five
profiles overlay each other indicating small within beam
angular divergence.
IV. CONCLUSIONS
Angular precision and integrity is essential to meet current
and future CMOS device performance requirements. The
VIISta 80 serial high current implanter integrated with its
automatic ion beam angle measurement and control system
provides state-of-the-art precision unmatched by any other
high current implanter. Process integrity requires closed-loop
ion beam incident angle control that this system provides.
Classical open-loop ion beam set-up methods do not account
for typical beam optical element changes and tuning variations
that effect beam incident angle and parallelism. The data
shown in this paper demonstrate the advanced process
performance capabilities of the VIISta 80.
1E+20
Center
Bottom
0.2
0.4
Right
Left
Top
1E+19
1E+18
1E+17
1E+16
1E+15
a.
b.
1E+14
0.0
Fig.9 B+, 70 keV 5E13 implant results a.) 0° incident angle- “angle
corrected” RS=814.5, σ =0.80% b.) -2° incident angle- “non-angle
corrected” RS=875.4, σ =1.3%
0.6
Depth (Microns)
Fig.11 SIMS profiles across wafer
0.8
1.0
ACKNOWLEDGEMENTS
The authors are extremely grateful for the help provided by
Sue Balbo in the preparation of this paper.
REFERENCES
[1]
[2]
Ukyo Jeong, Sandeep Mehta, Chris Campbell," Effects of Beam
Incident Angle Control on NMOS Source/Drain Extension
Applications" IIT 2002.
J.C. Olson, A. Renau, J. Buff, “Scanned Beam Uniformity Control in
the VIISta 810 Ion Implanter,” 1998 International Conference on Ion
Implantation Technology Proceedings, Kyoto, Japan, J. Matsuo, G.
Takaoka and I. Yamada, eds., Piscataway, NJ:OEEE 1999, pp. 169-172.