 |
Claims  |
|
|
What is claimed is:
1. A method for measuring a structure comprising multiple thin, metallic,
rectangular-shaped regions, each having a width of less than 5 microns and
being disposed between neighboring regions comprising a second,
non-metallic material, comprising:
exciting acoustic modes in at least one metallic, rectangular-shaped region
by irradiating the region with a spatially periodic excitation field
defined by a wavevector;
detecting the acoustic modes by diffracting a probe laser beam off a
modulated optical or physical property induced in the regions by the
acoustic modes to form a signal beam; and
analyzing the diffracted signal beam to determine a property of the
structure.
2. The method of claim 1, wherein the exciting further comprises
irradiating multiple metallic, rectangular-shaped regions with the
excitation field, and the detecting step further comprises diffracting a
probe laser beam off the surface ripple induced in each region by the
acoustic modes.
3. The method of claim 1, wherein the analyzing further comprises
determining the thickness of the metallic, rectangular-shaped region.
4. The method of claim 3, further comprising determining a thickness of the
metallic, rectangular-shaped region by analyzing a density and acoustic
properties of the metal comprised in the region, the wavevector, and a
frequency of the acoustic mode.
5. The method of claim 3, wherein the determining further comprises
analyzing a width of the metallic, rectangular-shaped region and a
distance separating the region and a neighboring, non-conducting region to
determine the thickness of the metallic, rectangular-shaped region.
6. The method of claim 1, further comprising determining a thickness of an
overlying or underlying film in the rectangular shaped region.
7. The method of claim 6, further comprising analyzing the probe beam
diffracted by the metallic, rectangular-shaped regions.
8. The method of claim 1, further comprising analyzing the signal beam to
determine a width of the metallic, rectangular-shaped region or a distance
separating consecutive metallic, rectangular-shaped regions.
9. The method of claim 1, wherein each of the metallic, rectangular-shaped
regions comprises copper, tungsten, aluminum, or alloys thereof.
10. The method of claim 9, wherein each of the metallic, rectangular-shaped
regions have a width of less than 1 micron.
11. The method of claim 9, wherein the metallic, rectangular-shaped region
comprises copper or an alloy thereof.
12. A method for measuring a structure comprising multiple thin, metallic,
rectangular-shaped regions, each having a width of less than 1 micron and
being disposed between neighboring regions comprising a second,
non-metallic material, comprising:
exciting acoustic modes in multiple metallic, rectangular-shaped regions by
simultaneously irradiating the regions with a spatially periodic
excitation field defined by a wavevector;
detecting the acoustic modes by diffracting a probe laser beam off a ripple
morphology induced in each of the regions by the acoustic modes; and
analyzing the signal beam to determine an average thickness of the
metallic, rectangular-shaped regions irradiated by the excitation field.
13. An apparatus for measuring a structure comprising multiple thin,
metallic, rectangular-shaped regions, each having a width of less than 5
microns and being disposed between neighboring regions comprising a
second, non-metallic material, comprising:
at least one excitation laser beam aligned to irradiate the region with a
spatially periodic excitation field defined by a wavevector to excite
acoustic modes in at least one metallic, rectangular-shaped region;
a probe laser beam that detects the acoustic modes by diffracting off a
modulated optical or physical property induced in the regions by the
acoustic modes to form a signal beam; and
an analyzer that analyzes the diffracted signal beam to determine a
property of the structure,
wherein the analyzer is further configured to analyze the probe beam
diffracted or reflected by the metallic, rectangular-based regions.
14. The apparatus of claim 13, wherein the excitation field is aligned to
irradiate multiple metallic, rectangular-shaped regions.
15. The apparatus of claim 14, wherein the probe laser beam is aligned to
diffract off the surface ripple induced in each region by the acoustic
modes.
16. The apparatus of claim 13, wherein the analyzer comprises a computer
configured to determine the thickness of the metallic, rectangular-shaped
regions.
17. The apparatus of claim 16, wherein the analyzer is configured to
determine a thickness of an overlying or underlying film in the
rectangular-shaped region.
18. The apparatus of claim 13, further comprising a photodetector
positioned to detect a diffracted or reflected beam to determine a width
of the metallic, rectangular-shaped region or a distance separating
consecutive metallic, rectangular-shaped regions.
19. The apparatus of claim 13, wherein each of the metallic,
rectangular-shaped regions comprises copper, tungsten, aluminum, or alloys
thereof.
20. The apparatus of claim 19, wherein each of the metallic,
rectangular-shaped regions have a width of less than 1 micron.
21. The apparatus of claim 13, wherein the excitation laser beam comprises
at least one optical pulse having a duration less than 1 nanosecond.
22. A method for measuring a thickness of a structure comprising multiple
thin, metallic, rectangular-shaped regions, each having a width of less
than 5 microns and being disposed between neighboring regions comprising a
second, non-metallic material, comprising:
measuring an acoustic property in a region of the structure;
analyzing the acoustic property of the structure to determine a mass in the
region of the structure; and
calculating a thickness of the structure from the mass and a periodicity of
the thin, metallic, rectangular-shaped regions.
23. A method for measuring a structure comprising a plurality of
polygon-shaped metal portions, comprising the steps of:
generating an excitation field;
aligning the excitation field so that it simultaneously irradiates at least
two of the polygon-shaped metal portions in a first region to excite an
acoustic component;
detecting the acoustic component by irradiating it with a probe laser beam
to generate a signal beam;
analyzing a feature of the signal beam corresponding to the acoustic
component to determine a property of the structure in the first region;
translating the structure or the excitation field and probe laser beam so
that the excitation field simultaneously irradiates at least two of the
polygon-shaped metal portions in a second region to excite an acoustic
component; and
repeating the excitation, detection, and analyzing steps to determine a
property of the structure in a second region to determine
position-dependent properties of the structure.
24. The method of claim 23, wherein the structure is comprised within a
single device comprising a semiconductor wafer.
25. The method of claim 24, wherein the first and second regions comprise
an respective arrays of polygon-shaped metal portions.
26. The method of claim 25, wherein the property measured in the first and
second regions is thickness.
27. The method of claim 24, wherein the polygon-shaped portion comprises
copper, and the property measured from the first and second regions is a
thickness of the copper.
28. The method of claim 24, wherein the first and second regions comprise
copper damascene structures, and the property measured from these regions
is a thickness of the copper in the damascene structure.
29. The method of claim 23, wherein the structure is a semiconductor wafer,
and the first region comprises a first array of polygon-shaped metal
portions, and the second region comprises a second array of polygon-shaped
metal portions.
30. The method of claim 29, wherein the property measured in the first and
second regions is thickness.
31. The method of claim 30, wherein the rectangular portion comprises
copper, and the property measured from the first and second regions is a
thickness of the copper.
32. The method of claim 31, wherein the first and second regions comprise
copper damascene structures, and the property measured from these regions
is a thickness of the copper in the damascene structure.
33. The method of claim 23, wherein the polygon-shaped metal portions are
substantially rectangular. |
|
 |
|
 |
Claims |
|
|
|
Description |
|
|
BACKGROUND
This invention relates to methods for determining the thickness of thin
films on small areas of structures used in microelectronics fabrication,
e.g., near a semiconductor wafer's edge or on a damascene-type structure.
During the fabrication of microelectronic devices, thin films of metals and
metal alloys are deposited on silicon wafers and for use as electrical
conductors, adhesion-promoting layers, and diffusion barriers.
Microprocessors, for example, use metal films of copper, tungsten, and
aluminum as electrical conductors and interconnects, titanium and tantalum
as adhesion-promoting layers, and titanium:nitride and tantalum:nitride as
diffusion barriers. Thickness variations in these films can modify their
electrical and mechanical properties, thereby affecting the performance of
the microprocessor. The target thickness values of these films vary
depending on their function: conductors and interconnects are typically
3000-10000 angstroms thick, while adhesion-promoting and diffusion-barrier
layers are typically between 100-500 angstroms thick. The deposition of
each of these films must be controlled such that the film's thickness is
within a few percent (e.g., 5-100 angstroms, a value roughly equivalent to
one or two seconds human fingernail growth) of its target value.
Furthermore, the uniformity of the film over the surface of the wafer must
be closely controlled in order to assure uniform behavior of the
individual microprocessors and, consequently, high manufacturing yields.
Because of these rigid tolerances, film thickness is often measured as a
quality-control parameter during and/or after the microprocessor's
fabrication.
The metal films are often deposited and patterned in complex geometries and
this complicates the measurement process. In a typical fabrication
process, a titanium:nitride film is deposited over the entire surface of a
silicon wafer. A tungsten film is then deposited onto the titanium:nitride
film to leave an "edge-exclusion zone", i.e., a small (about 1 or 2 mm)
region where the titanium:nitride is exposed, near the wafer's edge. The
edge-exclusion zone prevents delamination of the tungsten film near its
edges. Near this region, the thickness of the tungsten film rapidly
increases to its target value; this takes place over a distance of a few
hundred microns. Without this rapid increase in film thickness, devices
patterned near the wafer's edge-exclusion zone will contain non-ideal
tungsten films not having adequate thickness, and they will not meet
specifications.
An example of a complicated film geometry recently introduced in commercial
microelectronics fabrication is a "damascene" or "dual damascene"
structure. These structures, used especially to form copper conductors and
interconnects, are typically formed by a multi-step process: copper is
first deposited onto a wafer having a dielectric layer that has been
etched to have a series of trenches; the wafer is then polished by
chemical-mechanical polishing (CMP) to remove excess copper, leaving only
copper-filled trenches. The resulting structure is typically a series of
separated copper lines having a thickness of a few thousand angstroms, a
width of about 0.5 microns, a periodicity of about 2 microns, and a length
of several millimeters.
Measuring film thickness in and near the edge-exclusion zone and in
damascene-type structures is difficult and impractical using conventional
techniques. For example, blanket metal films are typically measured using
a 4-point probe. Here, two separated pair of conducting probes contact the
film; electrical resistance, as measured by the probes, relates to the
film's thickness. Because the spatial resolution of the 4-point probe is
typically a few hundred millimeters, this instrument is impractical for
both edge-profile and damascene-type structures. Moreover, a film's
resistance often depends on both its thickness and geometry, a
complication that further reduces the accuracy of the 4-point probe when
used to measure complex geometries. Another film-measuring instrument,
called a stylus profilometer, drags a stylus needle over a sample,
recording variations in topography. This instrument, however, is slow,
cumbersome, sensitive to slight amounts of sample curvature, and
inaccurate when used to measure relatively long distances (e.g., the
hundreds of microns required for tungsten build-up near the exclusion
zone).
In addition to these disadvantages, both 4-point probes and stylus
profilometers require contacting and thus contaminating a sample. These
instruments are therefore typically used on monitor or test samples,
rather than samples containing actual product. Other methods for measuring
the thickness of metal films, such as X-ray fluorescence and Rutherford
backscattering, are non-contact, but are slow and have poor spatial
resolution.
SUMMARY
In general, in one aspect, the invention provides a method for measuring a
structure that contains overlying and underlying films in a region where
the overlying film's thickness rapidly decreases until the underlying film
is exposed (e.g., an edge-exclusion structure). The method includes the
steps of: (1) exciting acoustic modes in a first portion of the region
with at least one excitation laser beam; (2) detecting the acoustic modes
with a probe laser beam that is either reflected or diffracted to generate
a signal beam; (3) analyzing the signal beam to determine a property of
the structure (e.g., the thickness of the overlying layer) in the first
portion of the region; (4) translating the structure or the excitation and
probe laser beams; and (5) repeating the exciting, detecting, and
analyzing steps to determine a property of the structure in a second
portion of the region.
In one embodiment, the exciting, detecting, analyzing, and translating
steps are repeated to determine a property of the structure in multiple
portions of the region. In one case, the above-mentioned steps are
repeated in an edge-exclusion structure until the thickness of the
overlying film is measured from where the underlying film is exposed to
where the overlying film's thickness is at least 80% of its average value.
This particular method can be extended so that the steps are repeated in
the structure until a diameter of the overlying film is measured.
Typically, this "diameter scan" embodiment includes repeating the
above-mentioned step on each side of the overlying film's diameter, and
measuring multiple points near the center of the film.
In another embodiment, the exciting, detecting, analyzing, and translating
steps are repeated until a property of the underlying film (e.g., the
width of the edge-exclusion zone) is measured from where it is exposed to
the edge of the wafer.
In typical embodiments: the overlying film is selected from a metal such as
tungsten, copper, aluminum, and alloys thereof; the underlying film is
selected from materials such as oxides, polymers, and metals such as
titanium, titanium:nitride, tantalum, tantalum:nitride, and alloys
thereof. These films are usually deposited on a silicon wafer.
The structure is typically measured using an optical method where the
acoustic modes are excited with at least one optical pulse having a
duration less than 1 nanosecond. In a particular embodiment, the exciting
step features exciting time-dependent acoustic modes in the structure by
directing a spatially periodic excitation radiation field defined by a
wavevector onto the sample. The radiation field, for example, is formed by
overlapping two optical pulses in time and space in or on top of the
sample. The detecting step then includes diffracting probe radiation off a
modulated optical or mechanical property induced on the sample's surface
by the acoustic modes. To determine thickness of the overlying layer, the
density and acoustic properties of the overlying and underlying layers,
the wavevector, and a frequency of the acoustic mode are analyzed (e.g.,
by comparing them to a mathematical model).
In another aspect, the invention features a method for measuring a
structure comprising multiple thin, metallic, rectangular-shaped or linear
regions, each having a width of less than 5 microns and being disposed
between neighboring regions that include a second, non-metallic material
(e.g., a damascene-type structure). The method includes the steps of: (1)
exciting acoustic modes in at least one metallic, rectangular-shaped
region by irradiating the region with a spatially periodic excitation
field defined by a wavevector; (2) detecting the acoustic modes by
diffracting a probe laser beam off a ripple morphology induced in the
regions by the acoustic modes; and (3) analyzing the diffracted signal
beam to determine a property of the structure (e.g., the thickness of the
metallic, rectangular-shaped regions).
In a particular embodiment, the exciting step includes irradiating multiple
metallic, rectangular-shaped regions with the excitation field. A probe
laser beam is then diffracted off the surface ripple induced in each
region by the acoustic modes. Thickness can be determined by analyzing a
density and acoustic properties of the metal included in the region, the
wavevector, and a frequency of the acoustic mode. Here, the width of the
metallic, rectangular-shaped region or a distance separating consecutive
regions may be used in the analysis. In still other embodiments, the
signal beam can be further analyzed (e.g., by monitoring diffraction of
the probe beam) to determine a width of the metallic, rectangular-shaped
region or a distance separating consecutive metallic, rectangular-shaped
regions.
In embodiments, each of the metallic, rectangular-shaped regions comprises
copper, tungsten, aluminum, or alloys thereof, and have a width of less
than 1 micron. The rectangular-shaped regions can also include more than
one layer. For example, the trench may be lined with tantalum and then
filled with copper.
In another aspect, the invention provides a method for measuring a
structure comprising multiple thin, metallic, rectangular-shaped regions,
each having a width of less than 1 micron and being disposed between
neighboring regions comprising a second, non-metallic material. The method
includes the steps of: (1) exciting acoustic modes in 30 multiple
metallic, rectangular-shaped regions by simultaneously irradiating the
regions with a spatially periodic excitation field defined by a
wavevector; (2) detecting the acoustic modes by diffracting a probe laser
beam off a modulated optical or physical property induced in each of the
regions by the acoustic modes; and (3) analyzing the signal beam to
determine an average thickness of the metallic, rectangular-shaped regions
irradiated by the excitation field.
The invention has many advantages. In particular, the method makes accurate
measurements of film thickness in and near the edge-exclusion zone, in
damascene-type structures, and in other small-scale structures.
Measurements feature all the advantages of optical metrology, e.g.,
noncontact, rapid and remote measurement over a small region. The method
collects data from a single measurement point having an area of between 10
and 100 microns in less than a few seconds. From these data film thickness
in the small-scale structures is determined with an accuracy and
repeatability of a few angstroms. For damascene-type structures, the
method simultaneously measures the thickness multiple metal lines lying
within the optical spot size with virtually no decrease in data quality,
accuracy, or repeatability. For typical films used in a microelectronic
device, the measurement determines thickness to within a fraction of a
percent of the film's true value.
In addition to thickness, the measurement determines the width of an
exclusion zone, the diameter of the useable area on the wafer, the film's
slope near the edge-exclusion zone, and properties of damascene-type
structures, such as the width and periodicity of the metal lines and the
number of defects in the structure.
The optical system used to make these measurements is compact, occupying a
footprint of about 2 square feet, and composed primarily of inexpensive,
commercially available components.
Because of its small size, the optical instrument an be a stand-alone unit,
or can be attached directly to a film-formation tool (e.g., a
chemical-vapor deposition tool, plasma-vapor deposition tool, a cluster
tool, or a vacuum chamber) or a film-processing tool (e.g., a
chemical-mechanical polisher). In these embodiments, the film-formation
tool includes an optical port (e.g., a glass window) that is transparent
to the excitation and probe radiation. Thus, during operation, the
film-measuring instrument is oriented so that the excitation and probe
radiation, and the diffraction signal, pass through the optical port.
Other features, aspects, and advantages of the invention follow from the
following detailed description, and from the claims.
BRIEF DESCRIPTION OF THE DRAWINGS
FIG. 1A is a schematic drawing showing an optical beam configuration of the
optical system according to the invention;
FIG. 1B is a top view of excitation and probe beams of FIG. 1A;
FIG. 2 is a plot (signal intensity vs. time) of a signal waveform measured
from a copper film using the optical beam configuration of FIG. 1A;
FIG. 3 is a flow chart describing a method for determining film thickness
by analyzing data taken from an edge-exclusion structure;
FIG. 4 is a cross-sectional, schematic drawing of an edge-exclusion
structure;
FIG. 5 is a plot of tungsten film thickness as a function of distance
measured using ISTS in an edge-exclusion structure;
FIG. 6 is a plot of tungsten film thickness as a function of distance
measured across a diameter of a silicon wafer structure using ISTS;
FIG. 7A is a top view of a damascene-type structure being measured with
ISTS;
FIG. 7B is a cross-sectional view of the damascene-type structure of FIG.
7A;
FIG. 8 is a plot (signal intensity vs. time) of a signal waveform measured
from a copper damascene-type structure similar to that shown in FIGS. 7A
and 7B;
FIG. 9 is a schematic drawing of an optical system and an optical detection
system for measuring a small-scale structure according to the method of
the invention;
FIG. 10 is a schematic drawing of an optical system, optical detection
system, computer, and signal processor for measuring a small-scale
structure according to the invention; and
FIG. 11 shows a flow chart for a computer-implemented algorithm for
determining the thickness in Damascene-type structures.
DETAILED DESCRIPTION
In FIGS. 1A, 1B, and 2, a thickness of a thin film 10 disposed on a
substrate 12 is measured in a small area 13 (e.g., in a damascene-type or
edge-profile structure) when two excitation laser pulses 16, 16' and a
probe laser pulse 18 irradiate the film. The excitation pulses 16, 16' are
short in duration (e.g. about 0.5 nanoseconds), have a wavelength that is
absorbed by the film, and are separated by an angle .alpha.. The probe
pulse 18 is relatively long (e.g. 30 several hundred nanoseconds or
longer) and has a wavelength that is not strongly absorbed by the film. In
this configuration, called a "four-wave mixing" geometry, the excitation
pulses 16, 16' overlap in time and space and interfere to form a spatially
and temporally varying excitation radiation field 14 in or on the surface
of the film 10. The field 14 is composed of a series of periodic,
sinusoidal "bright" regions 14a (i.e., constructive interference) and
"dark" regions 14b (i.e., destructive interference). The length and width
of the field 14, shown by the arrows 15a and 15b, are about 500 and 40
microns, respectively. When focused onto the film, the probe pulse 18
forms a second field 17 that is elliptical (roughly 90 microns by 25
microns) and lies completely within the excitation field 14.
The direction of the excitation field is defined by a wavevector that is
inversely proportional to the spatial distance between consecutive bright
(or dark) regions. The magnitude (q) of the wavevector is determined by
the angle .alpha. between the excitation pulses and the wavelength
.lambda..sub.1 of each pulse using the equation q=4.pi.sin(.alpha./2)
(.lambda..sub.1).sup.-1 =2.pi./.LAMBDA., where .LAMBDA. is the grating
wavelength.
The excitation radiation field 14 excites acoustic modes in the film 10
that have a wavelength and orientation corresponding to the excitation
wavevector. Excitation of the acoustic modes occurs via Impulsive
Stimulated Thermal Scattering ("ISTS"), a four-wave mixing technique that
is described in detail in U.S. Pat. No. 5,633,711 (entitled MEASUREMENT OF
MATERIAL PROPERTIES WITH OPTICALLY INDUCED PHONONS), U.S. Pat. No.
5,546,811 (entitled OPTICAL MEASUREMENT OF STRESS IN THIN FILM SAMPLES),
and U.S. Ser. No. 08/783,046 (entitled METHOD AND DEVICE FOR MEASURING
FILM THICKNESS, filed Jul. 15, 1996, the contents of which are
incorporated by reference. The acoustic modes induce a modulated optical
and physical property in the film (e.g., a time-dependent "ripple"
morphology and/or time-dependent refractive index change). In the case of
an induced time-dependent ripple morphology, this can be observed on the
film's surface. The frequency of the modulation depends on the thickness
of the film. Acoustic modes are detected by diffracting the probe pulse
off the modulated property to form at least two signal beams 20, 20'
disposed on each side of a reflected probe beam 18'. A photodetector
detects one or more of the signal beams to generate a signal waveform 30
similar to that shown in FIG. 2 which presents data taken from a nominal
3000 angstrom copper film. The Fourier transform of the signal waveform
30, indicated by the graph 35 in the figure inset, indicates the frequency
of the acoustic mode. To determine film thickness, the frequency is
analyzed along with the wavevector and the film's density and sound
velo | | |
