U.S. patent application number 12/938307 was filed with the patent office on 2011-02-24 for methods and apparatus for generating a library of spectra.
Invention is credited to Dominic J. Benvegnu, Ingemar Carlsson, Jeffrey Drue David, Sidney P. Huey, Lakshmanan Karuppiah, Harry Q. Lee, Jun Qian, Abraham Ravid, Boguslaw A. Swedek.
Application Number | 20110046918 12/938307 |
Document ID | / |
Family ID | 39793727 |
Filed Date | 2011-02-24 |
United States Patent
Application |
20110046918 |
Kind Code |
A1 |
Ravid; Abraham ; et
al. |
February 24, 2011 |
METHODS AND APPARATUS FOR GENERATING A LIBRARY OF SPECTRA
Abstract
A method of generating a library from a reference substrate for
use in processing product wafers is described. The method includes
measuring substrate characteristics at a plurality of well-defined
points of a reference substrate, measuring spectra at plurality of
measurement points of the reference substrate, there being more
measurement points than well-defined points, and associating
measured spectra with measured substrate characteristics.
Inventors: |
Ravid; Abraham; (Cupertino,
CA) ; Swedek; Boguslaw A.; (Cupertino, CA) ;
Benvegnu; Dominic J.; (La Honda, CA) ; David; Jeffrey
Drue; (San Jose, CA) ; Qian; Jun; (Sunnyvale,
CA) ; Huey; Sidney P.; (Fremont, CA) ;
Carlsson; Ingemar; (Milpitas, CA) ; Karuppiah;
Lakshmanan; (San Jose, CA) ; Lee; Harry Q.;
(Los Altos, CA) |
Correspondence
Address: |
FISH & RICHARDSON P.C.
P.O. BOX 1022
MINNEAPOLIS
MN
55440-1022
US
|
Family ID: |
39793727 |
Appl. No.: |
12/938307 |
Filed: |
November 2, 2010 |
Related U.S. Patent Documents
|
|
|
|
|
|
Application
Number |
Filing Date |
Patent Number |
|
|
12059435 |
Mar 31, 2008 |
7840375 |
|
|
12938307 |
|
|
|
|
60949498 |
Jul 12, 2007 |
|
|
|
60909639 |
Apr 2, 2007 |
|
|
|
Current U.S.
Class: |
702/158 ;
702/172 |
Current CPC
Class: |
G01N 21/31 20130101;
H01L 21/67161 20130101; H01L 21/67253 20130101; G01N 21/253
20130101; H01L 21/67766 20130101; H01L 21/67207 20130101 |
Class at
Publication: |
702/158 ;
702/172 |
International
Class: |
G06F 15/00 20060101
G06F015/00; G01B 11/06 20060101 G01B011/06; G01B 11/14 20060101
G01B011/14 |
Claims
1. A method of generating a library for use in processing product
wafers, the method comprising: measuring substrate layer thickness
at a first well-defined point and a second well-defined point of a
reference substrate with a first metrology system; measuring a
spectrum at a measurement point of the reference substrate with a
second monitoring system other than the first metrology system;
determining a closer well-defined point of the first well-defined
point and the second well-defined point to the measurement point;
and storing an association of the spectrum with the substrate layer
thickness of the closer well-defined point.
2. The method of claim 1, wherein substrate layer thickness is
measured prior to measuring the spectrum.
3. The method of claim 1, wherein measuring the spectrum comprises
scanning a sensor across the reference substrate and measuring a
plurality of spectra at a plurality of measurement points including
the spectrum at the measurement point.
4. The method of claim 3, further comprising measuring substrate
layer thicknesses of the reference substrate at a plurality of
well-defined points with the first metrology system, the plurality
of well-defined points including the first well-defined point and
the second well-defined point.
5. The method of claim 4, further comprising determining a closest
well-defined point of the plurality of well-defined points for each
of the plurality of measurement points.
6. The method of claim 5, further comprising, for each spectrum
from the plurality of spectra, storing an association of the
spectrum from the plurality of spectra with a substrate layer
thickness of the closest well-defined point.
7. The method of claim 6, further comprising: scanning a product
substrate other than the reference substrate with an optical
monitoring system to generate a measured spectrum of the product
substrate; determining a best matching spectrum from the plurality
of spectra to the measured spectrum of the product substrate; and
determining a substrate layer thickness associated with the best
matching spectrum.
8. The method of claim 7, wherein second monitoring system is the
optical monitoring system.
9. The method of claim 1, wherein the first well defined point and
the second well defined point are in different dies on the
reference substrate.
10. The method of claim 9, wherein the first well defined point and
the second well defined point are at the same relative position
within the different dies.
11. A computer program product, tangibly stored on machine readable
storage device, for generating a library for use in processing
product wafers, the product comprising instructions operable to
cause a processor to: cause substrate layer thickness to be
measured at a first well-defined point and a second well-defined
point of a reference substrate with a first metrology system; cause
a spectrum to be measured at a first measurement point of the
reference substrate with a second monitoring system other than the
first metrology system; determine a closer well-defined point of
the first well-defined point and the second well-defined point to
the first measurement point; and store an association of the
spectrum with the substrate layer thickness of the closer
well-defined point.
12. A method of generating a library for use in processing product
wafers, the method comprising: measuring a first value of a
substrate layer characteristic at a first well-defined point of a
reference substrate with a first metrology system and measuring a
second value of the substrate layer characteristic at a second
well-defined point of the reference substrate with the first
metrology system; measuring a spectrum at a measurement point of
the reference substrate with a second monitoring system other than
the first metrology system; determining a first distance from the
measurement point to the first well-defined point and a second
distance from the measurement point to the second well-defined
point; calculating a third value from the first value, the second
value, the first distance and the second distance; and storing an
association of the spectrum with the third value.
13. The method of claim 12, wherein the substrate layer
characteristic comprises a layer thickness.
14. The method of claim 12, wherein calculating the third value
comprises calculating a weighted average of the first value and the
second value with weighting based on the first distance and the
second distance.
15. The method of claim 12, further comprising measuring the
substrate layer characteristic of the reference substrate at a
plurality of well-defined points with the first metrology system to
generate a plurality of values, the plurality of well-defined
points including the first well-defined point and the second
well-defined point.
16. The method of claim 14, wherein the first well-defined point
and the second well defined point are the closest well-defined
points of the plurality of well-defined points to the measurement
point.
17. The method of claim 15, wherein measuring the spectrum
comprises scanning a sensor across the reference substrate and
measuring a plurality of spectra at a plurality of measurement
points including the spectrum at the measurement point.
18. The method of claim 16, further comprising, for each spectrum
from the plurality of spectra, determining distances from the
measurement point to two of the plurality of well-defined points
and calculating a value from values of the substrate layer
characteristic at the two of the plurality of well-defined points
and the distances.
19. The method of claim 17, further comprising: scanning a product
substrate other than the reference substrate with an optical
monitoring system to generate a measured spectrum of the product
substrate; determining a best matching spectrum from the plurality
of spectra to the measured spectrum of the product substrate; and
determining the value of the substrate layer characteristic
associated with the best matching spectrum.
20. The method of claim 12, wherein the first well defined point
and the second well defined point are in different dies on the
reference substrate.
21. The method of claim 19, wherein the first well defined point
and the second well defined point are at the same relative position
within the different dies.
22. A computer program product, tangibly stored on machine readable
storage device, for generating a library for use in processing
product wafers, the product comprising instructions operable to
cause a processor to: cause a first value of a substrate layer
characteristic to be measured at a first well-defined point of a
reference substrate with a first metrology system and cause a
second value of the substrate layer characteristic to be measured
at a second well-defined point of the reference substrate with the
first metrology system; cause a spectrum to be measured at a
measurement point of the reference substrate with a second
monitoring system other than the first metrology system; determine
a first distance from the measurement point to the first
well-defined point and a second distance from the measurement point
to the second well-defined point; calculate a third value from the
first value, the second value, the first distance and the second
distance; and store an association of the spectrum with the third
value.
Description
CROSS-REFERENCE TO RELATED APPLICATIONS
[0001] This application claims priority under 35 USC .sctn.119(e)
to U.S. patent application Ser. No. 12/059,435, filed on Mar. 31,
2008, which claims the benefit of prior U.S. Provisional
Application 60/949,498, filed Jul. 12, 2007, and U.S. Provisional
Application 60/909,639, filed Apr. 2, 2007 the entire contents of
which are hereby incorporated by reference.
TECHNICAL FIELD
[0002] This invention relates to metrology, and in one aspect to
optical monitoring of substrates during a chemical mechanical
polishing process.
BACKGROUND
[0003] An integrated circuit is typically formed on a substrate by
the sequential deposition of conductive, semiconductive, or
insulative layers on a silicon wafer. One fabrication step involves
depositing a filler layer over a non-planar surface and planarizing
the filler layer. For certain applications, the filler layer is
planarized until the top surface of a patterned layer is exposed. A
conductive filler layer, for example, can be deposited on a
patterned insulative layer to fill the trenches or holes in the
insulative layer. After planarization, the portions of the
conductive layer remaining between the raised pattern of the
insulative layer form vias, plugs, and lines that provide
conductive paths between thin film circuits on the substrate. For
other applications, such as oxide polishing, the filler layer is
planarized until a predetermined thickness is left over the non
planar surface. In addition, planarization of the substrate surface
is usually required for photolithography.
[0004] Chemical mechanical polishing (CMP) is one accepted method
of planarization. This planarization method typically requires that
the substrate be mounted on a carrier or polishing head. The
exposed surface of the substrate is placed against a rotating
polishing pad. The polishing pad may be either a "standard" pad or
a fixed-abrasive pad. The carrier head provides a controllable
load, i.e., pressure, on the substrate to push it against the
polishing pad. A polishing liquid, such as a slurry with abrasive
particles, is supplied to the surface of the polishing pad.
[0005] In order to determine the effectiveness of a polishing
operation, a "blank" substrate (e.g., a wafer with multiple layers
but no pattern) or a test substrate (e.g., a wafer with the pattern
to be used for device wafers) is polished in a tool/process
qualification step. After polishing, the substrate is removed from
the polishing system and the remaining layer thickness (or another
substrate property relevant to circuit operation, such as
conductivity) is measured at several points on the substrate
surface using an in-line or stand-alone metrology station. The
variation in layer thickness provide a measure of the wafer surface
uniformity, and a measure of the relative polishing rates in
different regions of the substrate. The in-line or stand-alone
metrology station can provide extremely accurate and reliable
thickness measurements (e.g., using ellipsometry) and precise
positioning of a sensor to desired measurement locations on the
substrate. However, this metrology process can be time-consuming,
and the metrology equipment can be costly.
[0006] One problem in CMP is determining whether the polishing
process is complete (i.e., whether a substrate layer has been
planarized to a desired flatness or thickness). Variations in the
initial thickness of the substrate layer, the slurry composition,
the polishing pad condition, the relative speed between the
polishing pad and the substrate, and the load on the substrate can
cause variations in the material removal rate. These variations
cause variations in the time needed to reach the polishing
endpoint. Therefore, for some applications, determining the
polishing endpoint merely as a function of polishing time can lead
to unacceptable variations in the post-polishing thickness of the
substrate layer. However, removal of the substrate from the
polishing apparatus for transportation to an in-line or stand-alone
metrology station can lead to an unacceptable reduction in
throughput.
[0007] Several methods have been developed for in-situ polishing
endpoint detection. One class of methods involve optically
monitoring the substrate during polishing, e.g., using an optical
sensor positioned in the platen that directs a light beam through a
window onto the substrate. However, measurements using such an
in-situ system usually cannot be precisely positioned at a desired
measurement location due to the motion of the substrate relative to
the sensor, and the measurements can be less accurate due to noise
generated by the polishing environment (e.g., absorption of light
by slurry), the limited time available for measurements, and the
need for real-time processing of the sensor data.
SUMMARY
[0008] This invention relates to a method of generating a library
from a reference substrate for use in processing product wafers.
The method includes measuring substrate characteristics a plurality
of well-defined points of a reference substrate, measuring spectra
at plurality of measurement points of the reference substrate,
there being more measurement points than well-defined points, and
associating measured spectra with measured substrate
characteristics.
[0009] Implementations of the invention may include one or more of
the following. Coordinates of the well-defined points and
coordinates of the measurement points may be stored. Associating
measured spectra with measured substrate characteristics can
include comparing coordinates of the well-defined points with
coordinates of the measurement points. Comparing coordinates of the
well-defined points with coordinates of the measurement points can
include determining a distance a spectra and a well-defined
point.
[0010] Associating measured spectra with measured substrate
characteristics can include determining a well-defined point that
is nearest to a particular measurement point, and associating the
substrate characteristic of the determined well-defined point with
the spectra of the particular measurement point. The substrate
characteristic can include a layer thickness, such as a pre- or
post-polish layer thickness. Identical spectra exhibiting different
layer thickness values can be removed. The plurality of
well-defined points can be at substantially similar relative
locations within different dies on the reference substrate. At
least some of the measurement points are spatially different than
the well-defined points. The substrate characteristics can be
measured prior to or after measuring the spectra. Measuring the
spectra can include scanning a sensor across the reference
substrate. A method of monitoring a substrate can include
generating a library from a reference substrate according to the
method above, scanning a product substrate with a optical
monitoring system to generate a plurality of spectra, and
determining substrate characteristics for the product substrate
based on the library. Scanning the product substrate can include
scanning with an in-situ monitoring system or scanning with an
in-line monitoring system.
[0011] In another aspect, a method of generating a library for use
in processing product wafers includes measuring a substrate layer
thickness at a first well-defined point and a second well-defined
point of a reference substrate, measuring a spectra at a first
measurement point of the reference substrate, determining the
closer of the first well-defined point and the second well-defined
point to the first measurement point, and associating the spectra
with the substrate layer thickness of the closer well-defined
point.
[0012] In another aspect, a computer program product, tangibly
stored on machine readable medium, includes instructions operable
to cause a processor to perform or cause the steps of the various
methods above.
[0013] In another aspect, a substrate processing system includes a
processing module to process a substrate, a factory interface
module configured to accommodate at least one cassette for holding
the substrate, a spectrographic monitoring system positioned in or
adjoining the factory interface module, and a substrate handler to
transfer the substrate between the at least one cassette, the
spectrographic monitoring system and the processing module.
[0014] Implementations of the invention may include one or more of
the following. The spectrographic monitoring system include may an
optical probe and may be configured to measure spectra at a
plurality of positions on the substrate while the substrate is
moving relative to the optical probe. The substrate may be moved by
the substrate handler and the optical probe may remain stationary.
Spectra may be measured in a plurality of positions that span a
diameter of the substrate in less than ten seconds. The plurality
of positions may form a non-linear path on the substrate, e.g., a
figure-eight path. The spectrographic monitoring system may include
an optical probe and may be configured to measure spectra at a
plurality of positions on the substrate without aligning the
optical probe to well-defined locations on the substrate. The
spectrographic monitoring system may be positioned in the factory
interface module. A notch alignment system may position a notch of
the substrate in a determined orientation.
[0015] As used in the instant specification, the term substrate can
include, for example, a product substrate (e.g., which includes
multiple memory or processor dies), a test substrate, a bare
substrate, and a gating substrate. The substrate can be at various
stages of integrated circuit fabrication, e.g., the substrate can
be a bare wafer, or it can include one or more deposited and/or
patterned layers. The term substrate can include circular disks and
rectangular sheets.
[0016] Possible advantages of implementations of the invention can
include one or more of the following. A library of spectra can be
assembled, and the spectra can be associated with physical
properties of the substrate. Spectra-based endpoint determination
can be made in-situ with greater speed and accuracy, and variations
in the post-polishing thickness of the substrate layer can be
reduced. Spectra-based measurements of substrate characteristics
can be made by in-line monitoring systems with great speed.
[0017] The details of one or more embodiments of the invention are
set forth in the accompanying drawings and the description below.
Other features, objects, and advantages of the invention will be
apparent from the description and drawings, and from the
claims.
DESCRIPTION OF DRAWINGS
[0018] FIG. 1 is a cross-sectional side view of an exemplary
chemical mechanical polishing apparatus having an in-situ optical
monitoring system.
[0019] FIG. 2 illustrates an exemplary path of spectra measurements
by an in-situ monitoring system across a substrate.
[0020] FIG. 3 shows an exemplary process for generating a library
that associates substrate characteristics with spectra.
[0021] FIG. 4 illustrates a portion of a reference wafer having
exemplary well-defined points.
[0022] FIG. 5 illustrates a data structure associating a substrate
characteristic with a coordinate for each well-defined points.
[0023] FIG. 6 illustrates a portion of a reference wafer having
exemplary measurement points.
[0024] FIG. 7 illustrates a data structure associating a spectrum
with a coordinate for each measurement point.
[0025] FIG. 8 illustrates a library with a data structure
associating spectra with substrate characteristics.
[0026] FIG. 9 illustrates an exemplary method for associating
spectra with substrate characteristics.
[0027] FIG. 10 illustrates an exemplary verification process for
data stored in a library.
[0028] FIG. 11 shows a method for using spectrum based endpoint
determination to determine an endpoint of a polishing step.
[0029] FIG. 12 is a top view of an exemplary substrate processing
system having an in-line spectrographic monitoring system.
[0030] FIG. 13 is a perspective view of an interior of an exemplary
factory interface module.
[0031] FIG. 14 is a side view of an exemplary factory interface
module having an in-line spectrographic monitoring system.
[0032] FIG. 15 illustrates an exemplary path of an optical probe of
the in-line spectrographic monitoring system across a reference
substrate during spectrographic measurements for library
generation.
[0033] FIG. 16 illustrates an exemplary path of an optical probe of
the in-line spectrographic monitoring system across a device
substrate during spectrographic measurements for data collection
for processing control.
[0034] Like reference symbols in the various drawings indicate like
elements.
DETAILED DESCRIPTION
[0035] Referring to FIG. 1, one or more substrates 10 will be
polished at a polishing station of a chemical mechanical polishing
(CMP) apparatus 20. A description of a polishing apparatus can be
found in U.S. Pat. No. 5,738,574, the entire disclosure of which is
incorporated herein by reference.
[0036] The polishing station includes a rotatable platen 24 on
which is placed a polishing pad 30. The platen 24 can be connected
to a platen drive motor (not shown). For most polishing processes,
the platen drive motor rotates platen 24 at thirty to two hundred
revolutions per minute, although lower or higher rotational speeds
may be used. The polishing station can also include a pad
conditioner apparatus to maintain the condition of the polishing
pad.
[0037] Polishing pad 30 typically has a backing layer 32 which
abuts the surface of platen 24 and a covering layer 34 which is
used to polish the wafer 10. Covering layer 34 is typically harder
than backing layer 32. However, some pads have only a covering
layer and no backing layer. Covering layer 34 can be composed of a
polyurethane with pores, e.g., a foamed polyurethane or cast
polyurethane with microspheres, and a grooved surface. Backing
layer 32 can be composed of compressed felt fibers leached with
urethane. A two-layer polishing pad, with the covering layer
composed of IC-1000 and the backing layer composed of SUBA-4, is
available from Rodel, Inc., of Newark, Del. (IC-1000 and SUBA-4 are
product names of Rodel, Inc.).
[0038] A carrier head 80 can be supported by a rotatable multi-head
carousel. Generally, the carrier head holds the wafer against the
polishing pad, distributes a downward pressure across the back
surface of the wafer, transfers torque from the drive shaft to the
wafer, and ensures that the wafer does not slip out from beneath
the carrier head during polishing operations. A description of a
carrier head can be found in U.S. Patent Publication No.
2006-0154580, the entire disclosure of which is incorporated herein
by reference. In addition, the carrier head 80 can be configured to
laterally oscillate across the polishing pad, e.g., move along a
radius of the polishing pad.
[0039] A polishing liquid, e.g., a slurry 38 containing abrasive
particles, can be supplied to the surface of polishing pad 30 by a
slurry supply port or combined slurry/rinse arm 39.
[0040] In typical operation, the platen is rotated about its
central axis 25, and the carrier head 80 is rotated about its
central axis 81 and translated laterally across the surface of the
polishing pad.
[0041] The polishing apparatus 20 also includes an in-situ optical
monitoring system 40, which can be used to determine a polishing
endpoint of the wafer being polished, as will be discussed below.
The optical monitoring system includes a light source 44 and a
light detector 46. Light passes from the light source 44, through
an optical access 36 in the polishing pad 30, impinges and is
reflected from the substrate 10 back through the optical access 36,
and travels to the light detector 46.
[0042] The optical access 36 through the polishing pad 30 to the
substrate can be provided by an aperture in the pad or a solid
window. The solid window can be secured to the polishing pad,
although in some implementations the solid window can be supported
on the platen 24 and project into an aperture in the polishing pad.
If the optical access 36 is in the form of a solid window, the
solid window can include, for example, a rigid crystalline or
glassy material (e.g., quartz or glass), a softer plastic material
(e.g., silicone, polyurethane or a halogenated polymer such as a
fluoropolymer), or a combination of these materials. The solid
window can be transparent to white light or light(s) at other
wavelengths.
[0043] A bifurcated optical cable 54 can be used to transmit the
light from the light source 44 to the optical access 36 and back
from the optical access 36 to the light detector 46. The bifurcated
optical cable 54 can include a "trunk" 55 and two "branches" 56 and
58.
[0044] The in-situ optical monitoring system 40 can include an
optical assembly 53 that is removably secured to the platen 24 in a
recess 26 in the platen 24 so that the optical assembly 53 rotates
with the platen 24. The optical access 36 can be aligned with the
recess 26 and the optical assembly 53. The recess 26 and the
optical access 36 can be positioned such that they have a view of
the substrate 10 during a portion of the platen's rotation,
regardless of the translational position of the carrier head. The
optical assembly 53 can hold one end of the trunk 55 of the
bifurcated fiber optic cable 54, which is configured to convey
light to and from a substrate surface being polished. The optical
head 53 can include one or more lenses to focus or collimate the
light beam. The optical head 53 can also include a window overlying
the end of the bifurcated fiber optic cable 54. Alternatively, the
optical assembly 53 can merely hold the end of the trunk 55
adjacent the solid window in the polishing pad. A refractive index
gel can be applied to a bottom surface of the window so as to
provide a medium for light to travel from the truck of the fiber
optic cable to the window.
[0045] The in-situ optical monitoring system 40 can also include an
in-situ monitoring module 50 that is removably secured to the
platen 24. The in-situ monitoring module 50 can include one or more
of the following: the light source 44, the light detector 46, and
circuitry for sending and receiving signals to and from the light
source 44 and light detector 46. For example, the output of the
detector 46 can be a digital electronic signal that passes through
a rotary coupler, e.g., a slip ring, in the drive shaft 22 to the
controller for the optical monitoring system. Similarly, the light
source can be turned on or off in response to control commands in
digital electronic signals that pass from the controller through
the rotary coupler to the module 50.
[0046] The in-situ monitoring module can also hold the respective
ends of the branch portions 56 and 58 of the bifurcated optical
fiber 54. The light source 44 is operable to transmit light, which
is conveyed through the branch 56 and out the end of the trunk 55
located in the optical head 53, and which impinges on a substrate
being polished. Light reflected from the substrate is received at
the end of the trunk 55 located in the optical head 53 and conveyed
through the branch 58 to the light detector 46.
[0047] In one implementation, the bifurcated fiber cable 54 is a
bundle of optical fibers. The bundle includes a first group of
optical fibers and a second group of optical fibers. An optical
fiber in the first group is connected to convey light from the
light source 44 to a substrate surface being polished. An optical
fiber in the second group is connected to received light reflecting
from the substrate surface being polished and convey the received
light to a light detector. The optical fibers can be arranged so
that the optical fibers in the second group form an X-like shape
that is centered on the longitudinal axis of the bifurcated optical
fiber 54 (as viewed in a cross section of the bifurcated fiber
cable 54). Alternatively, other arrangements can be implemented.
For example, the optical fibers in the second group can form V-like
shapes that are mirror images of each other. A suitable bifurcated
optical fiber is available from Verity Instruments, Inc. of
Carrollton, Tex.
[0048] There is usually an optimal distance between the polishing
pad window and the end of the trunk 55 of bifurcated fiber cable 54
proximate to the polishing pad window. The distance can be
empirically determined and is affected by, for example, the
reflectivity of the window, the shape of the light beam emitted
from the bifurcated fiber cable, and the distance to the substrate
being monitored. In one implementation, the bifurcated fiber cable
is situated so that the end proximate to the window is as close as
possible to the bottom of the window without actually touching the
window. With this implementation, the polishing apparatus 20 can
include a mechanism, e.g., as part of the optical assembly 53, that
is operable to adjust the distance between the end of the
bifurcated fiber cable 54 and the bottom surface of the polishing
pad window. Alternatively, the proximate end of the bifurcated
fiber cable is embedded in the window.
[0049] The light source 44 is operable to emit a broad wavelength
band of light, e.g., white light. In some implementations, the
white light emitted includes light having wavelengths of 200-800
nanometers. A suitable light source is a xenon lamp or a
xenon-mercury lamp. In some implementations, the light source
generates infrared or ultraviolet light.
[0050] The light detector 46 can be a spectrometer. A spectrometer
is basically an optical instrument for measuring properties of
light, for example, intensity, over a portion of the
electromagnetic spectrum. A suitable spectrometer is a grating
spectrometer. Typical output for a spectrometer is the intensity of
the light as a function of wavelength.
[0051] Optionally, the in-situ monitoring module 50 and optical
assembly 53 can include additional other sensor elements in
addition to the spectrometer, such as an eddy current sensor, a
monochromatic interferometric optical sensor, or a friction
sensor.
[0052] The light source 44 and light detector 46 are connected to a
computing device 48 operable to control their operation and to
receive their signals. The computing device can include a
microprocessor situated near the polishing apparatus, e.g., a
programmable computer, such as a personal computer. The computing
device can, for example, synchronize activation of the light source
44 with the rotation of the platen 24.
[0053] As shown in FIG. 2, the optical monitoring system can make a
sequence of spectral measurements as the optical assembly 53 and
optical access 36 scan across the substrate. Each of points 201-211
represent a location on the substrate 10 where light from the
in-situ monitoring system impinges and reflects off to provide a
spectral measurement. As shown in FIG. 2, the locations can trace
an arc across the substrate due to the rotation of the platen 24.
Optionally, the computer can cause the light source 44 to emit a
series of flashes starting just before and ending just after the
substrate 10 passes over the optical access 36 module, with each
flash corresponding to a measurement location. Alternatively, the
computer can cause the light source 44 to emit light continuously
starting just before and ending just after the substrate 10 passes
over the in-situ monitoring module.
[0054] The computing device 48 can be programmed to store spectral
intensity measurements from the detector, to display the spectra on
an output device, to calculate the remaining thickness, amount
removed, and polishing rate from the spectral intensity
measurements, and/or to detect the polishing endpoint. The
computing device 48 also can be configured to cause, for example,
the polishing rate and polishing time of the polishing apparatus to
be adjusted based upon the received light.
[0055] Generally, in order to calculate a thickness of a layer on
the substrate or to detect a polishing endpoint based on the
spectrum measured by the optical monitoring system 40, a measured
spectrum is compared to a library of reference spectra.
[0056] FIG. 3 shows an exemplary process 300 for generating a
library that associates reference spectra with substrate
characteristics.
[0057] Initially, at least one characteristic of a reference
substrate, e.g., of a substrate layer, is measured at multiple
locations on the reference substrate (step 302). For each location,
the measured characteristic and the location of the measurement are
stored, e.g., in a first data structure in a computer-readable
medium.
[0058] The reference substrate should have the same pattern and die
feature geometry as an actual product substrate would have at the
same point in the manufacturing process, although the reference
substrate need not itself be intended to be a product substrate.
The characteristic should be measured for at least a substrate that
has approximately the thickness as the product substrate will have
when measured by a spectrographic system that will use the library.
For example, if the product substrate will be measured by an
in-line system pre or post-polishing, then the reference substrate
should be measured with approximately the expected pre or
post-polishing thickness, respectively. If the product substrate
will be measured by an in-situ monitoring system, then the
reference substrate should be measured for at least the desired
post-polishing substrate layer thickness, but as discussed below,
the characteristic can be measured for one or more reference
substrates at multiple different stages of polishing of the
substrate layer.
[0059] The characteristic can be a physical property of the
substrate that impacts the performance of circuitry on the
substrate. An exemplary physical characteristic is a thickness of a
film of interest, e.g., the outermost layer undergoing processing.
Other thickness-derived characteristics can include step height or
erosion. Other possible physical characteristics of the film
include conductivity. Alternatively, the characteristic can be a
manufacturing metric, e.g., a yield. In addition, the film of
interest need not be the outermost layer, e.g., the physical
characteristic can be a thickness of an underlying layer.
[0060] The substrate characteristics can be measured using a
metrology system that provides precise positioning of a sensor to a
desired measurement location on the substrate. The metrology system
can be part of an in-line or stand-alone metrology station. The
metrology station can include positional sensors and alignment
mechanism for aligning the substrate and the sensors so that the
same location is repeatedly and accurately measured for different
substrates. If the metrology system measures substrate layer
thickness, it can be a non-contact optical metrology system, such
an optical metrology system that uses spectral intensity and/or
polarization information to calculate layer thickness, or it can be
a contact profilometer. If the metrology system measures substrate
layer conductivity, it can include a four-point probe. Suitable
optical metrology systems for measuring the substrate layer
thickness are available from Nova Measuring Instruments and
Nanometrics.
[0061] The characteristic is measured at a multiple locations of
interest on the reference substrate. In some implementations, these
locations are "well-defined" points, i.e., locations at which a
metrology device can generate an accurate and reliable measurement
without relying on this invention. For example, in the context of a
conventional non-contact optical metrology device, a well defined
location is a location at which the optical model used by the
metrology device can be used to accurately calculate the substrate
layer thickness a priori from the measured properties of the
reflected light (e.g., spectral intensity and polarization) with a
reasonable amount of computational processing power. For example,
in the context of a conventional four-point probe, a well defined
location is a location with sufficiently large conductive area for
placement of the probe. Locations having a lower density of
geometrical features than other discrete regions of the wafer can
be selected as well-defined points. For example, well-defined
points may include regions in which bond pads are placed, or
regions in which surfaces of uniform material composition are
formed.
[0062] The well-defined points can be selected so that each
measurement on a particular substrate occurs for locations in
different dies but at the same relative position within each die.
For a particular substrate at a particular stage of polishing, the
number of locations measured can be equal to or less than, e.g.,
less than, the number of dies on the substrate. The measurement
locations can be selected to be generally uniformly spaced across
the substrate.
[0063] FIG. 4 illustrates a reference wafer having exemplary
well-defined points. Referring to FIG. 4, the reference wafer 406
may contain one or more die features 402 (exemplary dies are
labeled D.sub.1, D.sub.2 . . . D.sub.m-1 and D.sub.m). To provide
accurate thickness profile analysis of the reference wafer 406, a
thickness from each well-defined point 402 (exemplary well-defined
points are labeled WP.sub.1, WP.sub.2 . . . WP.sub.m-1 and
WP.sub.m) is measured. Specifically, light is impinged upon each
well-defined point, as shown by the measurement spot 404, and
portions of the light reflected off the well-defined points
300a-300f are received. Based on spectra detected in the reflected
light, thickness measurement at these well-defined points 300a-300f
can be obtained.
[0064] FIG. 5 illustrates a first data structure generated from
collected data that associates the coordinates of at least some of,
and possibly each, well-defined point with a corresponding
substrate characteristic. Referring to FIGS. 4 and 5, the substrate
characteristic, such as thicknesses, of the reference wafer is
measured at well-defined points WP.sub.1, WP.sub.2 . . . WP.sub.m-1
and WP.sub.m positioned at coordinates (Xw.sub.1, Yw.sub.1),
(Xw.sub.2, Yw.sub.2) . . . (Xw.sub.m-1, Yw.sub.m-1) and (Xw.sub.m,
Yw.sub.m), respectively. Of course, a different coordinate system
(e.g., R, .theta.) could be used.
[0065] As shown, wafer characteristic T.sub.1 is measured for
well-defined point WP.sub.1 at coordinates (Xw.sub.1, Yw.sub.1).
Similarly, wafer characteristics T.sub.2, . . . T.sub.m-1 and
T.sub.m are measured for well-defined points WP.sub.2 . . .
WP.sub.m-1 and WP.sub.m at coordinates (Xw.sub.2, Yw.sub.2), . . .
(Xw.sub.m-1, Yw.sub.m-1) and (Xw.sub.m, Yw.sub.m), respectively.
These measurements can then be stored in the first data structure.
If the substrate characteristics were calculated from measured
spectra, then the data structure can optionally also store the
measured spectrum associated with each coordinate. In addition, for
each measurement or group of measurements, the data structure can
store a unique identifier of the reference substrate, and data
indicating the stage of polishing of the reference substrate layer
(e.g., an elapsed polishing time or a number of platen
rotations).
[0066] In some implementations, substrate characteristic are
calculated for at least some intermediate points. These
intermediate points can have the same relative positioning within
each die as the well-defined points. The intermediate points can be
well-defined points at which the substrate characteristic was not
measured, but can also be other points in a die.
[0067] The substrate characteristic of the intermediate points can
be calculated by linear interpolation or extrapolation from
measured well-defined points, particularly the nearest several
measured well-defined points, e.g., nearest two to four
well-defined points, on the reference substrate. For example,
referring to FIG. 4, if substrate layer thicknesses T.sub.1 and
T.sub.m-1 are measured for points WP.sub.1 and WP.sub.m-1, and the
well-defined and intermediate points are uniformly spaced, then the
thickness for intermediate point IP can be calculated as the
average of T.sub.1 and T.sub.m-1. More generally, the linear
interpolation can be a weighted average of nearby measured
well-defined points with weighting based on relative distance to
the well-defined points.
[0068] Referring back to FIG. 3, at step 304, spectra are measured
at multiple locations across the reference substrate. For each
location, the spectra and the location of the measurement are
stored, e.g., in a data structure in a computer-readable
medium.
[0069] The spectra are measured for at least some locations
(hereinafter "measurement points") other than the well-defined
points, although it is permissible for spectra to also be measured
at locations that overlap with the well-defined points. However,
the measurement points need not selected so that each measurement
occurs at the same relative position within a die.
[0070] The spectra can be measured with an optical monitoring
system that does not provide precise positioning of a sensor to a
desired measurement location on the substrate. For example, the
spectra can be measured with an optical monitoring system that
scans a sensor across the substrate at relatively high speed (e.g.,
across a 300 mm diameter wafer in less than 10 seconds, e.g, in
less than 5 seconds), and without halting. The optical monitoring
system can be part of an in-situ monitoring system, e.g., at a
polishing station, or an in-line metrology station. The spectra can
measured using an optical monitoring system with substantially the
same configuration as the in-situ monitoring system to be used at
the polishing system (e.g., as described above with reference to
FIG. 1). In one implementation, the spectra are measured using the
same in-situ optical monitoring system as the one that will be used
in the polishing system. In another implementation, the monitoring
system can be an in-line or stand alone system that otherwise
mimics the in-situ monitoring system, e.g., using the same light
source, detector, sampling rate, fiber optic connector and window,
but is not in a polishing station.
[0071] For a particular substrate at a particular stage of
polishing, the number of measurement points can be greater than the
number of measured well-defined locations, and can be much greater,
e.g., ten or more times greater, e.g., one-hundred or more times
greater. At least some dies include more than one measurement
point. In general, the spacing between measurement points is less
than the spacing between the well-defined locations, and the
density of measurement points is also greater than the density of
the well-defined locations. The number of measurement points can be
greater than the number of dies on the substrate.
[0072] For example, referring to FIG. 6, assuming that the spectra
are measured using an in-situ optical monitoring system as
described with reference to FIG. 1 above, the light beam creates a
sweeping path 610 and spectra are measured along the sweeping path,
as indicated by the measurement points 612 MP.sub.1, MP.sub.2 . . .
MP.sub.m-1 and MP.sub.m.
[0073] The number of measurement points can depend on the sampling
rate of the detector 46. The detector 46 can have a sampling rate
between about 10 and 100 Hz, corresponding to a sampling period
between about 2.5 and 100 milliseconds. Each time the detector 46
is sample, the in-situ optical monitoring system 40 retrieves
spectral data, such as intensity and reflectance data, from an
associated measurement point 612. The computing device 48 can cause
the light source 44 to emit a series of light beam starting just
before and ending just after the reference wafer 406 passes over
the optical module 53, or the light beam can be on
continuously.
[0074] Although FIG. 6 shows only eleven measurement points
MP.sub.1, MP.sub.2, . . . MP.sub.10 and MP.sub.11, this is
illustrative and there could be many more measurement points. The
number of measurement points depends on the platen rotation rate
and the sampling rate of the detector 46. Of course, a lower
triggering rate can result in fewer (and more widely spaced)
measurement points, whereas a faster triggering rate can result in
a larger number of (and more closely spaced) measurement points.
Similarly, a lower rotation rate can result in a larger number of
measurement points, whereas a faster rotation rate can result in
fewer measurement points.
[0075] Also, more than a single sweep can be performed on a
particular reference substrate at a particular stage of polishing
to produce a measurement points. From the measurement points, the
computing device 48 accumulates a set of intensity or reflectance
measurements, each associated with a measurement time (e.g., time
between a previous sweep and a subsequent sweep).
[0076] Spectra from the measurement points 612 can be collected
using an optical monitoring tool capable of producing measurement
in broad wavelength range, covering, for example, the deep
ultraviolet (e.g., wavelengths below 300 nm), ultraviolet, visible
or infrared wavelength regions. The wavelength range in which
measurement is to be taken can include an entire or a partial
segment of the in-situ optical monitoring system's operating
wavelength range.
[0077] For illustrative purposes, spectra S.sub.1, S.sub.2, . . .
S.sub.8, S.sub.9 . . . S.sub.m-1 and S.sub.m are measured at
measurement points MP.sub.1, MP.sub.2 . . . , MP.sub.8, MP.sub.9 .
. . MP.sub.m-1 and MP.sub.m positioned at coordinates (Xm.sub.1,
Ym.sub.1), (Xm.sub.2, Ym.sub.2) . . . (Xm.sub.8, Ym.sub.8),
(Xm.sub.9, Ym.sub.9) . . . (Xm.sub.m-1, Ym.sub.m-1) and (Xm.sub.m,
Ym.sub.m), respectively.
[0078] FIG. 7 illustrates a second data structure generated from
collected data that associates the coordinates of each measurement
point with a corresponding spectrum. As shown, spectra S.sub.1 is
measured at coordinates (Xm.sub.1, Ym.sub.1). Similarly, spectra
S.sub.2 . . . S.sub.m-1 and S.sub.m are measured at coordinates
(Xm.sub.2, Ym.sub.2) . . . (Xm.sub.n-1, Ym.sub.n-1) and (Xm.sub.n,
Ym.sub.n) respectively. Of course, a different coordinate system
(e.g., R, .theta.) could be used.
[0079] The coordinate position of each measurement point at which a
spectrum is obtained can be determined by using methods similar to
those described in U.S. Pat. Nos. 7,018,271, 7,097,537, and
7,153,185 the disclosures of which is incorporated herein by
reference. In particular, these disclosures describe calculation of
a radial positions of a measurement, and an angular position can be
calculated from a carrier head angular position at the time of
measurement, e.g., as sensed by a rotary encoder. Of course, the R,
.theta. coordinate determination can be transformed into another
coordinate system (e.g., X, Y).
[0080] In addition, for each measurement or group of measurements,
the second data structure can store a unique identifier of the
reference substrate, and data indicating the stage of polishing of
the reference substrate layer (e.g., an elapsed polishing time or a
number of platen rotations).
[0081] Returning to FIG. 2, at step 206, spectra measured from
measurement points are associated with substrate characteristics
based on predetermined conditions. The associated spectra and
substrate characteristics are stored to form a library. For
example, each spectrum can be linked to a substrate characteristic
of a nearby well-defined point based on the coordinates of the
measurement point at which the spectrum was measured. Associating
spectra with substrate characteristics will be described in further
detail below with reference to FIG. 9.
[0082] FIG. 9 illustrates an exemplary process 900 for associating
spectra with wafer characteristics. A set of well-defined points
can be determined for use in generating the library (step 902).
Typically, for a particular substrate at a particular stage of
polishing, all of the well-defined points at which the substrate
characteristic was measured would be used, but it is possible for
fewer than all of the well-defined points to be used to generate
the library. Similarly, a set of measurement points is determined
for use in generating the library (step 904). Again, typically for
a particular substrate at a particular stage of polishing, all of
the measurement points at which spectra were measured would be
used, but it is possible for fewer than all of the measurement
points to be used to generate the library.
[0083] For each measurement point in the set, one of well-defined
points is selected, and the substrate characteristic of the
selected well-defined point is assigned to the spectra of the
measurement point in the library (step 904). The selected
well-defined point is near the measurement point, e.g., one of the
four closest measurement points. In one implementation, the
well-defined point closest to the measurement point is selected.
This can be accomplished by comparing the coordinates of the
measurement point to the coordinates of well-defined points and/or
calculating distances between the measurement point and the
well-defined points. Once the distance between a measurement point
and neighboring well-defined points are determined, an association
can be established by identifying a well-defined point closest to
the measurement point, and linking the spectrum previously measured
at that well-defined point to the wafer characteristic(s)
associated with the measurement point. In another implementation,
the selected well-defined point is the well-defined point in the
same die as the measurement point.
[0084] As an example, referring to FIG. 6, assuming that
coordinates (Xm.sub.1, Ym.sub.1) and (Xm.sub.2, Ym.sub.2) of
measurement points MP.sub.1 and MP.sub.2 are closest to
well-defined points WP.sub.1 and coordinates (Xm.sub.8, Ym.sub.8)
and (Xm.sub.9, Ym.sub.9) of measurement points MP.sub.8 and
MP.sub.9 are closest to well-defined points WP.sub.2, then
associations between spectra S.sub.1 and S.sub.2 and wafer
characteristic T.sub.1, and between spectra S.sub.8 and S.sub.9 and
wafer characteristic T.sub.2 are established (see FIG. 8). Of
course, associates between the spectra for the other measurement
points and substrate characteristics for other well-defined points
can also be made.
[0085] In some implementations, to expedite the process of distance
determination, a predetermined distance or zone from a well-defined
point can be identified in advance so that spectra measured at
measurement points falling within the predetermined distance or
zone are automatically recognized and associated with the wafer
characteristics at that well-defined point. For example, still
referring to FIG. 6, a spectrum of any measurement point falling
inside a first zone 602a is automatically associated with the wafer
characteristics of the well-defined point WP.sub.1, and spectrum of
any measurement point falling inside a second zone 602b is
automatically associated with the wafer characteristics of the
well-defined point WP.sub.2. The definition of the zone for each
well-defined point can be stored in the first data structure.
[0086] In these implementations, associations for spectra of
measurement points falling inside an overlapping region of both the
first and second zones can be established by using the distance
technique discussed above. For example, measurement point MP.sub.7
is situated between the boundaries of the first zone 602a and the
second zone 602b. If the distance between the measurement point
MP.sub.7 and the well-defined point WP.sub.1 is shorter than that
between the measurement points MP.sub.7 and the well-defined point
WP.sub.2, then the association between the spectrum at the
measurement point MP.sub.7 and substrate characteristics at the
well-defined point WP.sub.1 is established. Conversely, if the
distance between the measurement point MP.sub.7 and the
well-defined point WP.sub.1 is longer than that between the
measurement point MP.sub.7 and the well-defined point WP.sub.2,
then the association between the spectrum at measurement point
MP.sub.7 and substrate characteristics at the well-defined point
WP.sub.2 is established.
[0087] FIG. 8 illustrates a third data structure generated from
collected data that associates spectra with substrate
characteristics and that forms the library. As shown, spectrum
S.sub.1 is associated with thickness T.sub.1, spectrum S.sub.2 is
associated with thickness T.sub.1, spectrum S.sub.8 is associated
with thickness T.sub.2 and spectrum S.sub.9 is associated with
thickness T.sub.2. Optionally, information related to the distance
between each measurement point and well-defined point, including
coordinates thereof, can be stored in the library.
[0088] Returning to FIG. 2, at step 208, it is determined whether
spectra and substrate characteristic measurements of a reference
substrate are needed at additional different polishing stages. If
it is determined that measurements are needed at additional
different polishing stages ("Yes" branch of step 208), steps
202-206 are repeated. In general, steps 202-206 can be repeated
until spectra and substrate characteristics are accumulated for a
sufficient number of different thicknesses to ensure reliable
operation during polishing of actual product wafers.
[0089] In one implementation, the reference substrate is initially
measured at a partially polished state. After substrate
characteristics and spectra have been measured in, the reference
substrate can be transferred back to the polishing apparatus to
partially polish an additional incremental amount of substrate
layer material. In fact, spectra can be collected during the
polishing process (e.g., using the in-situ monitoring system
described above to collect spectra from the last platen rotation
before polishing is halted). The reference substrate is then
removed from the polishing apparatus for measurement of the
substrate characteristics at the well-defined locations, e.g.,
using a conventional in-line or stand-alone metrology system. Of
course, the reference substrate can then be sent back to the
polishing system for additional polishing.
[0090] Otherwise (at "No" branch of step 208), process 200
indicates that the library is prepared to be used for processing
actual product wafers (step 210).
[0091] Steps 202 and 204 can be performed in the order listed or in
reverse of the order listed. Thus, spectra measurement at multiple
measurement points across the reference substrate can be performed
before or after the measurement of substrate characteristics at
well-defined points. In addition, in some implementations, some
operations of steps 202-206 can be performed in another order or in
parallel to achieve the same result. For example, an association
between spectrum measurements and wafer characteristics can be
performed as each spectrum is received. As another example, if the
substrate characteristics are calculated for some of the
well-defined points (e.g., by linear interpolation), the
calculation can be performed after the closest well-defined point
has been identified for a spectrum.
[0092] The library can reside in the memory of the computing device
48. The library can be updated with new data (e.g., if a product
substrate is directed to a metrology station, then spectra from the
product substrate collected from the in-situ monitoring system,
e.g., from the last platen rotation before polishing was halted,
could be associated with the substrate characteristics measured at
the metrology station). If desired, the library also can include
spectra that are not collected but are theoretically generated.
Other parameters such as time in which the spectra are measured
also can be stored in the library. In addition, the library is not
limited to storing data collected from a single substrate, and can
include spectra collected from multiple substrates.
[0093] Because precise alignment of the measurement tools at the
well-defined points is no longer required, the library can
significantly increase the overall speed with which substrate
characteristics can be determined, and thus the throughput of the
polishing apparatus can be increased. To optimize the throughput of
the polishing apparatus, a high density of spectra and wafer
characteristics covering an entire wafer area are captured before,
during and after polishing so that a sufficient number of wafer
characteristics and spectra measurements is stored. This enables
high speed, high volume, precise real time thickness extraction and
reporting. However, if during polishing of an actual product wafer,
a measured spectrum is found not to have a matching spectrum stored
in the library, the library can be immediately updated to include
the measured spectrum and its associated wafer information.
[0094] Once a sufficient number of established associations are
identified and collected, the library can be used for monitoring
during processing of actual product wafers. During actual
processing, the optical monitoring system sweeps across a product
substrate and measures a sequence of spectra from the reflected
light, and the library can be searched for a matching spectra. The
search may include direct comparison of the measured spectra to
those stored in the library, or using a combination of searching
and fitting algorithms. The substrate characteristics associated
with the spectra selected from the search can then be used for
monitoring or control of the polishing process.
[0095] In some implementations, endpoint can be called when a
measured spectrum has a desired substrate characteristic. For
example, as discussed above, for the spectra measured during
polishing, the closest matching spectrum in the library can be
identified, e.g., using searching and/or fitting algorithms. If the
substrate characteristics, e.g., thickness, of the matching
spectrum in the library has the desired characteristic, e.g., a
desired thickness, then the polishing endpoint is triggered.
[0096] In another implementation, the library is searched in
advanced for a desired endpoint criterion, e.g., a desired
thickness, and one or more spectra which have a substrate
characteristic with the desired criterion are identified as desired
spectra. Then, during polishing, for the spectra measured during
polishing, the closest matching spectrum in the library can be
identified. Polishing can be halted when the measured spectrum
matches a desired spectra from the library.
[0097] In some implementations, the library is not used for
endpoint determination, but is merely used for monitoring and/or
feedback control of pressure applied by the carrier head to the
substrate. For example, endpoint could be detected using a
difference traces between the current spectra measured during
polishing and a reference spectrum, as described in U.S. Patent
Application Publication No. 2007/0042675, the disclosure of which
is incorporated herein by reference in its entirety.
[0098] Optionally, the spectra collected can be verified to enhance
the reliability of the library. FIG. 10 illustrates an exemplary
verification process 1000 for the library. Referring to FIG. 10,
the library is sorted (step 1000). This operation functions to
expedite the verification process of the data stored in the
library. If desired, this operation can be bypassed if the library
contains less than a predetermined number of data. Once the library
is sorted, spectra (and/or other wafer parameters) stored in the
library are analyzed (step 1004). If it is determined that two or
more spectra stored in the library are substantially identical yet
exhibit a different thickness (step 1006), then both spectra are
permanently discarded from the library ("Yes" branch of step 1008).
Otherwise ("No" branch of step 1008), the analysis step is
resumed.
[0099] In some implementations, spectra stored in the library are
normalized, averaged and/or filtered to enhance the reliability of
the library. For example, spectra matching can be performed after
processing and filtering the measured spectra (e.g., using high
pass filter or low pass filter) to remove noise and interference.
The spectra also can be compensated for optical system distortions
and other artifacts, or be matched to different optical response
used to collect the spectra for the library. This may include, for
example, intensity variations and wavelength dependent scattering
due to the feature structure, array dimensions, numerical aperture
effects, wavelength range and polarization.
[0100] In some implementations, each measured raw spectra can be
normalized to remove light reflections contributed by mediums other
than the film or films of interest. Normalization of spectra
facilitates the comparison process discussed above. Light
reflections contributed by media other than the film or films of
interest include light reflections from, for example, the polishing
transparent window 36 and from the base silicon layer of the wafer.
Contributions from, for example, a transparent window 36 can be
estimated by measuring the spectrum of light received by the in
situ optical monitoring system 40 under a dark condition (i.e.,
when no wafers are placed over the in situ optical monitoring
system 40). Contributions from, for example, the silicon layer can
be estimated by measuring the spectrum of light reflecting off a
bare silicon wafer. The contributions can be obtained prior to
commencement of the polishing step.
[0101] A measured raw spectrum can be normalized as follows:
normalized spectrum=(A-Dark)/(Si-Dark)
[0102] where A is the raw spectrum, Dark is the spectrum obtained
under the dark condition, and Si is the spectrum obtained from the
bare silicon wafer.
[0103] Optionally, the collected spectra can be sorted based on the
region of the pattern that has generated the spectrum, and spectra
from some regions can be excluded from the endpoint calculation. In
particular, spectra that are from light reflecting off scribe lines
can be removed from consideration. Different regions of a reference
wafer usually yield different spectra (even when the spectra were
obtained at a same point of time during polishing).
[0104] For example, a spectrum of the light reflecting off a scribe
line in a wafer can be different from the spectrum of the light
reflecting off an array of the wafer. Because of their different
shapes, use of spectra from both regions of the pattern usually
introduces error into the endpoint determination. However, the
spectra can be sorted based on their shapes into a group for scribe
lines and a group for arrays. Because there is often greater
variation in the spectra for scribe lines, usually these spectra
can be excluded from consideration to enhance precision.
[0105] A high pass filter also can be applied to the measured raw
spectra. Application of the high pass filter can remove low
frequency distortion of the average of the subset of spectra. The
high pass filter can be applied to the raw spectra, their average,
or to both the raw spectra and their average.
[0106] In some implementations, based on the current spectra of
each zone and the variations thereof, the computing device 48 can
determine the flatness of the wafer and the polishing uniformity
for CMP tool and process qualification. For example, the computing
device 48 can applies process control and endpoint detection logic
to determine when to change process and polish parameter and to
detect the polishing endpoint. Possible process control and
endpoint criteria for the detector logic include local minima or
maxima, changes in slope, threshold values in amplitude or slope,
or combinations thereof. The spectra of light reflected from a
wafer can be frequently monitored and collected as polishing
progresses. Based on the reflected spectra, the computing device 48
can determine an endpoint of a polishing process.
[0107] If more than one current spectra is measured for a platen
revolution, then the spectra can be grouped, combined, e.g.,
averaged within each group, and the averages are designated to be
current spectra. The spectra can be grouped by radial distance from
the center of the wafer. By way of example, for a given platen
rotation, a first current spectrum can be obtained, e.g., by
averaging, from spectra measured as points 211 and 219 (FIG. 3), a
second current spectrum can be obtained from spectra measured at
points 212 and 218, a third current spectra can be obtained from
spectra measured at points 213 and 217, and so forth.
[0108] FIG. 11 shows another method 1100 for determining an
endpoint of a polishing step. Initially, index values are assigned
to the spectra in the library (step 1104). The index values can be
selected to monotonically increase as polishing progresses, e.g.,
an index values can be proportional to a number of platen
rotations. Thus, each index number can be a whole number, and the
index number can represent the expected platen rotation at which
the associated spectrum would appear. The library can be
implemented in memory of the computing device of the polishing
apparatus.
[0109] A wafer from the batch of wafers is polished, and the
following steps are performed for each platen revolution. One or
more spectra are measured to obtain a current spectra for a current
platen revolution (step 1106). The spectra are obtained as
described above. The spectra stored in the library which best fits
the current spectra is determined (step 1108). The index of the
library spectrum determined to best fits the current spectra is
appended to an endpoint index trace (step 1110). Endpoint is called
when the endpoint trace reaches a reference index, e.g., the index
of a spectrum having the desired thickness or other substrate
characteristic (step 1112).
[0110] Although implementations for determining a film thickness
have been described, other parameters including shallow trench
depth, step height of various semiconductor materials (e.g.,
silicon dioxide, silicon nitride), an area of trench or active
region of the wafer, or thickness of silicon dioxide or pad
layers.
[0111] Although the discussion above focuses on use of the library
in a polishing endpoint detection system, the library could also be
used in for an in-line spectrographic metrology system, e.g., an
in-line system that scans a sensor across the substrate at
relatively high speed. This in-line metrology system could be used
before or after processing, e.g., polishing, of the substrate, and
the substrate characteristics derived from the measured could be
used for feed-forward or feed-back control of the polishing system.
For example, if the library associates thicknesses with spectra,
then the in-line metrology system could measure substrate layer
thickness at multiple points along a radius or diameter of the
substrate prior to polishing, and the measured layer thickness data
could be used to control the polishing system (e.g., select
endpoint criteria or polishing head pressures) during polishing of
that substrate. As another example, the in-line metrology system
could measure substrate layer thickness at multiple points along a
radius or diameter of the substrate after polishing, and the
measured layer thickness data could be used to control the
polishing system (e.g., select endpoint criteria or polishing head
pressures) during polishing of a subsequent substrate. Due to the
large number of spectra stored in the system, the system can
provide reliable measurements of the substrate characteristics
without precise positioning of the sensor to any well-defined
point, thereby permitting the measurements to be made at the
in-line station at high throughput.
[0112] An implementation of a substrate processing system 8 that
includes an in-line spectrographic metrology system 500 is
illustrated in FIG. 12. The substrate processing system 8 includes
the chemical mechanical polishing apparatus 20, a factory interface
module 100, a wet robot 140, and a cleaner 170. Substrates 10,
e.g., silicon wafers with one or more layers deposited thereon, are
transported to the substrate processing system 8 in cassettes 12,
and are extracted from the cassettes 12 by the factory interface
module 100 for transport to the polishing apparatus 20 and the
cleaner 170. The operations of the substrate processing system 8
are coordinated by controller 48, such as one or more programmable
digital computers executing control software. Some of the modules,
such as the wet robot 140 and cleaner 170, could be omitted,
depending on the configuration of the processing system, and the
processing system could include other modules, such as a deposition
or etching apparatus.
[0113] The polishing apparatus 20 can includes a series of
polishing stations 150 and a transfer station 152. The transfer
station 152 serves multiple functions, including receiving
individual substrates 10 from the wet robot 140, washing the
substrates and loading the substrates into carrier heads. Each
polishing station can includes a rotatable platen holding a
polishing pad 30. Different polishing pads can be used at different
polishing stations. A rotatable carousel 154 that holds four
carrier heads 80 is supported above the polishing stations (drive
systems above the carrier heads and the carrier head over the
transfer station are not illustrated in FIG. 12 to provide a
clearer top view). The carousel 154 rotates to carry the substrates
between the polishing stations 150 and the transfer station
152.
[0114] The cleaner 170 can be generally rectangular shaped cabinet
with a front wall 171, a back wall 172, and two side walls 174. The
interior of the cleaner 170 is divided into an input or staging
section 176 and a cleaning section 178. The staging section 176
includes a substrate-pass through support 180 and an indexable
buffer 182, each of which can hold one or more substrates in a
vertical orientation. The cleaner also includes a walking beam 184
which can hold a substrate in a vertical orientation.
[0115] The wet robot 140 is configured to transport the substrate
between the staging section 176 and the polishing apparatus 20.
[0116] The factory interface module 100 can be substantially
rectangular in shape and include an outer wall 101, an inner wall
102, a first side wall 104, and a second side wall 106. The outer
wall 101 can be aligned with a cleanroom wall. A plurality (e.g.,
four) cassette support plates 110 project from the outer wall 101
into the cleanroom to accept the cassettes 12, and a plurality of
cassette ports 112 are formed in the outer wall 101 to permit
transport of the substrates from the cassettes 12 into the factory
interface module 100. The inner wall 104 mates against a front wall
171 of the cleaner 170 and shares an entry port 120 (to the staging
section 176) and an exit port 122 (from the end of the cleaning
section 178) with the cleaner front wall 171. The inner wall 102
and the cleaner front wall 170 may be combined into one structure,
and there may be additional ports from the factory interface module
100 to the cleaner 170.
[0117] One or more factory interface wafer handlers 130
(hereinafter simply "robot"), depicted in greater detail in FIGS.
13 and 14, are housed within the factory interface module 100. In
some implementations the factory interface robot 130 has a base
132, a rotatable vertical shaft 134 extending from the base 132, a
horizontally extendible articulated arm 136 supported by the shaft
134, a rotary actuator 138 at the end of the articulated arm 136,
and a substrate gripper 139 (in phantom below the substrate 10 in
FIG. 13) supported by the rotary actuator 138. The vertical shaft
134 is capable of lifting and lowering the articulated arm 136
vertically. Rotation of the vertical shaft 134 permits rotary
motion of the articulated arm 136 about a vertical axis, and the
articulated arm 136 is configured to extend and retract
horizontally. The rotary actuator can be pivotally connected to the
end of the articulated arm 136 so as to be rotatable about a
vertical axis. In addition, the rotary actuator 138 can rotate the
substrate gripper 139 about a horizontal axis. The factory
interface robot 130 thus provides a wide range of motion to
manipulate the substrate held by the gripper 139. The gripper 139
can be a vacuum chuck, an electrostatic chuck, an edge clamp, or
similar wafer gripping mechanism. The factory interface robot can
also include an optical detector to sense whether a substrate is
being held by the gripper 140. Sensors, e.g., encoders, can be used
to detect the position of the movable elements of the robot 130 so
that the position of the gripper 139 and substrate 10 can be
calculated.
[0118] The base 132 can be supported on a linear rail 131 that
extends parallel to the inner and outer walls 102, 100. A motor can
drive the factory interface robot 130 laterally along the rail 131
to access the entry port 120, the exit port 122, the cassette ports
112 (FIG. 12 illustrates two positions along the slide 142 for the
factory interface robot 130), and the in-line spectrographic
metrology system 500 within the factory interface 100.
[0119] As shown in FIG. 14, the in-line spectrographic monitoring
system 500 operates similarly to the in-situ optical monitoring
system, and includes a light source 44 and a light detector 46.
Light passes from the light source 44, through an optical guide,
impinges and is reflected from a substrate 10 held in the factory
interface 100, back through the optical guide, and travels to the
light detector 46. As with the in-situ system, a bifurcated optical
cable 54 can be used to transmit the light from the light source 44
to the substrate 10 and back from the substrate 10 to the light
detector 46. The bifurcated optical cable 54 can include a "trunk"
55 with an end 504 fixed in a position selected to be in proximity
to substrate when the substrate is to be scanned by the metrology
system, and two "branches" 56 and 58 connected to the light source
44 and light detector 46, respectively. The light source 44 and
light detector 46 are connected to a computing device 48 that
performs the various computational steps in the metrology process.
Although FIG. 14 illustrates the light source 44 and a light
detector 46 as positioned outside the factory interface 100, these
components could be located inside the factory interface 100.
[0120] A bracket 502 secured to a wall of the factory interface 100
can hold the trunk 55 of the optical fiber 54 in a fixed position
inside the factory interface 100. The robot 130 can be controlled
to sweep the substrate at a working distance of two to thirty-five
millimeters from the end 504 of the optical fiber.
[0121] The factory interface 100 can also include a pre-aligner 510
to position the substrate in a known rotational position. The
pre-aligner 510 includes a rotatable support 512, such as a
pedestal, possibly with a vacuum or electrostatic chuck, an edge
support ring, or support pins, onto which the substrate can be
placed. In addition, the pre-aligner 510 includes a notch detection
system, such as an optical interrupter sensor 520, to sense when
the substrate notch is at a specific angular position. During
creation of a library, the reference substrate is placed by the
robot 130 on the support 512, the support 512 rotates so that the
sensor 520 detects the substrate notch, and rotates to place the
substrate notch in a predetermined angular orientation. Then the
robot 130 retrieves the substrate from the support 512. Thus,
substrates which might be in an uncertain angular position, e.g.,
after a polishing operation, have a known orientation when scanned
by the in-line spectrographic monitoring system 500, thus
permitting accurate determination of the x-y (or r-.theta.)
position of the measurements on the substrate. Because the position
of the spectra measurements is known with higher accuracy, the
reliability of the association of spectra measurements with
substrate characteristics is improved.
[0122] The substrate processing system 8 can operate in two modes:
an initial library creation mode and a later in-line monitoring
mode. In the library creation mode the substrate processing system
can generate a library for a particular type of substrate, e.g., a
particular pattern and a particular metal or dielectric level in
the fabrication process. In general, a separate library is created
for each different metal or dielectric level in the fabrication
process for each different pattern. In the in-line monitoring mode,
the substrate processing system 8 uses the previously generated
library to perform quickly determine the characteristics of
substrates undergoing processing based on the measured
spectrographic data.
[0123] Library generation occurs generally as discussed above with
respect to FIG. 3. A reference substrate with a particular pattern
and at a particular point in the fabrication process is measured
using a conventional metrology system that provides very precise
positioning of a sensor to well-defined locations on the substrate,
e.g., a Nova or Nanometrics optical metrology system. The
measurements can be made before or after a polishing step in the
fabrication process. At least one characteristic of a reference
substrate, e.g., layer thickness, is measured at multiple
well-defined locations on the reference substrate, and the measured
characteristic and the measurement location are stored, e.g., in a
first data structure. The metrology system can be an in-line system
within the processing system 8, or a stand alone system. However,
one potential advantage of using the in-line spectrographic
monitoring system described herein is that the processing system 8
need not include the conventional metrology system. In particular,
because the conventional metrology system is needed only for
accurate substrate characteristic measurements during library
generation (rather than during production), a single stand alone
metrology system should be able provide the necessary measurements
for library generation for multiple processing systems 8.
[0124] The reference substrate 10a is placed into a cassette 112
and extracted from the cassette into the factory interface by the
robot 130. The robot 130 moves the reference substrate to engage
the pre-aligner so that the position of the substrate can be
precisely identified. Then the reference substrate is held by the
robot and moved past the optical probe. A sequence of
spectrographic measurements are generated using the in-line
metrology system 500, the position of each spectrographic
measurement on substrate is determined, and the spectra and
measurement locations are stored, e.g., in a second data
structure.
[0125] For gathering of spectrographic data for library generation,
an exemplary path 520 of the optical probe 504 across a reference
substrate 10a having a notch 11 is illustrated in FIG. 15. The path
520 can include several arcs 522 that pass along the substrate
edge, e.g., within 8 mm, e.g., within 5 mm, of the substrate edge,
to ensure that a significant number of measurements are obtained
near the substrate edge.
[0126] Once both spectrographic data and characteristic
measurements at well-defined locations are obtained, the library
can then be generated by associating each spectrographic
measurement from the first data structure with a characteristic
measurement from the first data structure at a nearby well-defined
location. The spectra can be measured at different locations on the
reference substrate 10 by the in-line spectrographic monitoring
system 500 before or after the substrate characteristic is measured
by the metrology system.
[0127] Returning to FIGS. 13 and 14, during processing of device
substrates, e.g., in a normal polishing operation, an unpolished
substrate is retrieved by the factory interface robot 130 from one
of the cassettes 112. The factory interface robot 130 "picks" the
substrate, e.g., by vacuum suction, and transports the unpolished
substrate at relatively high speed past the optical probe of the
in-line spectrographic monitoring system 500 in the factory
interface. Thus, the robot 130 acts as the stage to hold the
substrate during the measurement process. The in-line
spectrographic monitoring system 500 measures spectra for a
sequence of points across the substrate as the substrate is
scanned, and a layer thickness measurement is generated for at
least some of the measured points. These pre-polish layer thickness
measurements can be used to adjust the polishing process parameters
for the substrate.
[0128] The robot 130 then transports the substrate through the
entry port 120 to the staging section 176. There, the substrate is
placed in either the pass-through support 180 or the indexible
buffer 182. The wet robot 140 then extracts the substrate 10 from
the staging section 176 and places the substrate 10 into the
transfer station 152 of the polishing apparatus 20. From the
transfer station 152, the substrate 10 is carried to one or more
polishing stations 150 to undergo chemical mechanical polishing.
After polishing, the wet robot 140 transports the substrate 10 from
the transfer station 152 to the walking beam 184 in the cleaner
120. The walking beam 184 transports the substrate through the
cleaner section 178 of the cleaner 120. While the substrate 10 is
transported through the cleaner section 178, slurry and other
contaminants that have accumulated on substrate surface during
polishing are removed.
[0129] The factory interface robot 130 removes the substrate 10
from the cleaner 120 through the exit port 122, and transports the
polished substrate at relatively high speed past the optical probe
of the in-line spectrographic monitoring system 500 in the factory
interface 100. Again, the in-line spectrographic monitoring system
500 measures spectra for a sequence of points across the substrate
as the substrate is scanned, and a layer thickness measurement is
generated for at least some of the measured points. These
post-polish layer thickness measurements can be used to adjust the
polishing process parameters for a subsequent substrate. Finally,
the factory interface robot 130 returns the substrate 10 to one of
the cassettes 112.
[0130] For gathering of spectrographic data during device substrate
processing for control of polishing parameters, an exemplary path
530 of the optical probe 504 across a device substrate 10b having a
notch 11 is illustrated in FIG. 16. In some implementations, the
path 530 describes a "figure eight" shape on the substrate. The
path 530 can include several arcs 532 that pass along the substrate
edge, e.g., within 8 mm, e.g., within 5 mm, of the substrate edge,
to ensure that a significant number of measurements are obtained
near the substrate edge.
[0131] The robot 130 can move the substrate at a fairly high speed
across the substrate. For example, the robot could move a 300 mm
diameter substrate to cause the optical probe to trace the path
shown in FIG. 16 in about three to seven seconds, e.g., about six
seconds. The detector 46 can have a sampling rate of about 130 to
150 samples per second, e.g., 142 samples per second (the light
source 44 can flash on for each spectrographic measurement). Thus,
assuming that path 530 is traced over about 6 seconds, about 850
spectra can be measured along the path. Due to the high speed of
the in-line measurement, e.g., a velocity of about 150-350 mm/sec
during many measurements, during production each and every
substrate can undergo both pre-polish and post-polish measurement
without impacting substrate throughput (for throughput <85 wafer
per hour). Thus, for each substrate, thickness measurements at a
variety of radial positions on the substrate can be used to control
processing conditions for that substrate or for a subsequent
substrate.
[0132] Optionally, the in-line spectrographic metrology system
could be housed in a separate module 160 connected to the factory
interface module 100. For example, one of the side walls 104 or 106
(side wall 106 in the implementation shown in FIG. 12) mates with a
wall 161 of the metrology module 160 and shares an access port 124.
The side wall 104 and the monitoring system wall 161 may be
combined into one structure, and there may be additional ports from
the factory interface module 100 to the metrology module 160. The
metrology module 160 could include a separate robot for the
substrate, or the factory interface robot 130 could manipulate the
substrate, to cause the substrate to be scanned past the
spectrographic probe.
[0133] The subject matter described herein contemplates a
comprehensive thin-film metrology and polishing system, which
combines measurements of patterned wafers irrespective of locations
of the measurements. It offers both real-time, in-line measurements
(i.e. performed within a semiconductor fabrication tool) and also
rapid multi-point (i.e. mapping) at-line measurements of film
thickness, composition, and electronic properties. The present
concepts can be applied broadly to many of the critical electronic
materials that are processed in semiconductor fabrication tools,
including polysilicon, silicon dioxide, silicon nitride, and other
dielectrics.
[0134] Implementations and all of the functional operations
described in this specification can be implemented in digital
electronic circuitry, or in computer software, firmware, or
hardware, including the structural means disclosed in this
specification and structural equivalents thereof, or in
combinations of them. Implementations described herein can be
implemented as one or more computer program products, i.e., one or
more computer programs tangibly embodied in an information carrier,
e.g., in a machine readable storage device or in a propagated
signal, for execution by, or to control the operation of, data
processing apparatus, e.g., a programmable processor, a computer,
or multiple processors or computers. A computer program (also known
as a program, software, software application, or code) can be
written in any form of programming language, including compiled or
interpreted languages, and it can be deployed in any form,
including as a stand alone program or as a module, component,
subroutine, or other unit suitable for use in a computing
environment. A computer program does not necessarily correspond to
a file. A program can be stored in a portion of a file that holds
other programs or data, in a single file dedicated to the program
in question, or in multiple coordinated files (e.g., files that
store one or more modules, sub programs, or portions of code). A
computer program can be deployed to be executed on one computer or
on multiple computers at one site or distributed across multiple
sites and interconnected by a communication network.
[0135] The processes and logic flows described in this
specification can be performed by one or more programmable
processors executing one or more computer programs to perform
functions by operating on input data and generating output. The
processes and logic flows can also be performed by, and apparatus
can also be implemented as, special purpose logic circuitry, e.g.,
an FPGA (field programmable gate array) or an ASIC (application
specific integrated circuit).
[0136] The above described polishing apparatus and methods can be
applied in a variety of polishing systems. Either the polishing
pad, or the carrier head, or both can move to provide relative
motion between the polishing surface and the wafer. For example,
the platen may orbit rather than rotate. The polishing pad can be a
circular (or some other shape) pad secured to the platen. Some
aspects of the endpoint detection system may be applicable to
linear polishing systems (e.g., where the polishing pad is a
continuous or a reel-to-reel belt that moves linearly). The
polishing layer can be a standard (for example, polyurethane with
or without fillers) polishing material, a soft material, or a
fixed-abrasive material. Terms of relative positioning are used; it
should be understood that the polishing surface and wafer can be
held in a vertical orientation or some other orientations.
[0137] Particular implementations have been described. Other
implementations are within the scope of the following claims. For
example, the actions recited in the claims can be performed in a
different order and still achieve desirable results.
* * * * *