 |
Description  |
|
|
BACKGROUND OF THE INVENTION
1. Technical Field
This invention relates generally to optical inspection instruments and,
more specifically, to such instruments that inspect a surface of interest,
such as a wafer, for example during in-line processing.
2. Discussion of the Related Art
There has been much investigation in the field of inspection systems. In
particular, inspection systems are needed in the semiconductor industry
for inspecting wafers. The significance for detecting and identifying
submicron surface defects on such wafers is due in part to the present
semiconductor industry move from 0.250 micron to 0.180 micron and, in less
than ten years, to 0.130 micron fabrication architectures. The latter
architecture is so exacting that it will require the detection of 2 nm
substrate defects and 40 nm sized particles on unpatterned silicon wafers.
In addition, the industry is scaling up from 200 mm to 300 mm diameter
wafers with fewer defects commercially permitted, and rapid detection
required at all processing stages. The current industry projection is that
about 9% of all wafer production will use the 300 mm format by the year
2002. To meet these needs, defect data must be processed in near-real-time
to allow correction of processing problems through, for example,
statistical process control (SPC) techniques.
Many surface roughness inspection systems are available. These include:
high resolution microscopes such as the atomic force microscope, optical
microscopes such as the phase contrast microscope, other optical
measurement systems such as ellipsometers, and mechanical contact methods
that use a stylus. For sub-micron resolution, most of these techniques are
not suitable for in-process surface inspection. High-resolution
microscopes require cumbersome surface preparations and expensive
operations. Optical microscopes, in general, do not have sufficient
resolution and accuracy. Ellipsometry or spectroscopy also do not provide
adequate surface roughness information. Mechanical stylus devices are
simply out of the question.
Among the possible methods is the optical heterodyne (frequency-shifted)
microscope. The heterodyne microscope is an interferometric microscope
where the signal beam is frequency-shifted relative to a reference beam.
With this method, the signal containing the optical phase information
(surface roughness) can be electronically detected by comparing the phase
of the beams from different portions of the wafer surface. The problems
with this method are: (a) a critical focusing requirement, (b) low
throughput rate due to slow scanning, (c) inadequate lateral resolution
(>0.2 mm), (d) limited sensitivity (<100 nm), and (e) inadequate
information on the surface deposit materials. In particular with regard to
problem (b), a serious drawback with microscopy, in general, is in
assessing the number of defects over a large wafer area by scanning with a
micron-sized area of view. This would require hours, if not days, to view
an entire wafer surface even with a multiple array of detectors.
Another candidate technology, namely scatterometry, has a drawback as
presently, generally implemented. A significant limitation with
scatterometry is that the intensity of scattering is typically measured at
oblique angles, excluding the specular beam. Under this condition, the
diameter of particles that can be realistically detected using an in-line
tool must be greater than 80 nm (i.e., in order to capture enough light to
detect particles, the particle size must be greater than about 80 nm.
U.S. Pat. No. 5,343,290 to Batchelder, et al describes a dark field
illumination heterodyne interferometer for particle detection. A
heterodyne interferometer is combined with a dark field illumination for
improved surface particle detection sensitivity. The U.S. Pat. No.
5,486,919 to Tsuji also describes an optical heterodyne interferometer to
detect defects on patterned wafers utilizing different states of
polarization for incident and scattered light. The Batchelder and Tsuji
heterodyne interferometers provide photon counting detection (Shot noise
limited) for the small scattered light from the particles, although the
grazing illumination described in both patents will only provide lower
sensitivity (only good for a larger particle detection). The problem with
these methods is that they cannot effectively discriminate particles on
the wafer surface, (which are desired to be detected) from particles
within the beam suspended in the air, (which are not desired to be
detected.)
U.S. Pat. No. 5,030,842 to Koshinaka, et al describes a method to detect
particles using two illumination beams, where the phase of one beam with
respect to the other is modulated so that the detected light scattered
from a particle in the illuminated area can be distinguished from all
other light. As they point out in their patent, however, it will detect
particles suspended in the air as long as they are within the overlapped
region of the two illumination beams. This will increase a chance of
producing a false signal.
U.S. Pat. No. 6,081,325 to Leslie, et al describes a sophisticated
scatterometer, which can detect defects and particles on unpatterned and
patterned wafer surfaces. The system uses several photo-multiplier tubes
and a charge-coupled detector (CCD) camera. A focused laser beam
illuminates the sample surface and the light scattered from a particle or
defect is monitored by these multiple numbers (4) of photo-multiplier
tubes. Utilizing the asymmetrical nature of scattering from particles or
defects, the system discriminates between defects (or particles) and the
wafer pattern. In addition they use a CCD camera and obtain an image of
the sample surface illuminated by the focused laser beam: light scattered
from only the illuminated spot is focused on a pixel of CCD camera. The
wafer is scanned multiple times to provide a desirable image. However,
this imaging method does not provide photon-counting detection. Also, the
rescanning of the surface is time consuming a particularly undesirable for
commercial applications.
The objective of the present invention is to overcome these shortcomings of
the existing devices and to achieve photon detection using a CCD camera,
and to completely discriminate between particles on a sample surface and
those suspended in air. This is achieved without using any external
frequency shift or phase modulation device and focused illumination beam
scanning.
SUMMARY OF THE INVENTION
In accordance with the present invention, an apparatus is provided for
inspecting a surface of an object. The apparatus includes the following
major parts: a light source, a photodetector, an optical assembly, a
positioning assembly, and a controller. The light source is configured to
generate an illumination light beam. The illumination beam, when incident
upon a particle on the surface, is scattered to define a scattered beam.
The optical assembly is configured to direct the scattered beam and a
reference beam, derived from the illumination beam, to the photodetector.
The positioning assembly is configured to move the object such that the
surface moves relative to the illumination beam. Finally, the controller
is coupled to the photodetector and is configured to detect the presence
of the particle in accordance with an interference pattern from a
superposition of the reference beam and the scattered beam.
The invention is suitably adapted for use in the inspection of
semiconductor wafers, using optical detection to holographically record
light scattered from a particle on a wafer surface illuminated with, in a
preferred embodiment, a laser or other coherent source. The invention
utilizes two beams, at least one of which is incident on a wafer surface.
These two beams are preferably derived from the same source and are
subsequently re-combined to form the above-mentioned interferometric
pattern. In one embodiment, a geometrical configuration is used such that
the illumination beam strikes the surface at normal or near normal
incidence. A specular reflection of the illumination beam from the wafer
surface is used as the reference beam and is interferometrically combined
with light backscattered (the "scattered beam") from any contaminant
particles on the surface. The device, when used in a preferred
environment, is especially effective for scanning unpatterned
semiconductor wafers, since they have only a relatively few, isolated
contaminant particles present.
In a preferred embodiment, the light source may comprise conventional
apparatus, such as a diode-pumped solid-state laser, a multiple wavelength
ion laser, or even a partially coherent source like a "white-light" Xenon
lamp. In operation, a wafer surface is illuminated by the illumination
beam. The scattered light from a deposit (particle) is detected as a
fringe pattern formed by the interference between the scattered light and
a reference beam generated from the same light source. The interference
pattern will be a spatially fluctuating component of the light power,
which can be detected using the photodetector.
In another preferred embodiment, the interference pattern is preserved
during scanning. This is accomplished by using Time Delay and Integration
(TDI). In particular, the preservation is accomplished by transporting the
wafer surface such that the modulated light signal (i.e., interference
pattern) moves across the face of the photodetector in synchronism with an
electronic position shift of the register portions of the photodetector.
Since the shifting of the image data in the registers is substantially
locked to the scanning speed of the wafer, the inspection apparatus
according to the invention is basically immune against background optical
noise, such as that caused by scattering particles in the air. In the
described embodiment, the apparatus may operate as a "photon-counter",
limited only by the quantum effect shot noise.
In yet another embodiment, composition analysis of detected particles is
performed. The composition analysis of such a contaminant particle is
carried out using a spectrometer. By analyzing the spectrum of the
interference pattern using predetermined data, the composition of the
contaminant can be identified.
Inasmuch as the amplitude of the scattered signal is proportional to the
volume of the contaminant particle, the controller can determine volume
density. The apparatus detects the scattered beam as a modulation pattern
caused by the interference between the reference beam and the scattered
beam. This interference pattern can be detected with, in a preferred
embodiment, a CCD camera.
In such embodiments employing a CCD camera, the CCD camera may be
configured in a variety of ways including, but not limited to, both a
linear scanning array and an area array. This latter configuration can
function as an integrating device or as a frame transfer camera. As an
integrating device, the CCD camera may be controlled by the controller to
operate as a TDI sensor. In this technique, the detector array registers
are clocked so that the charge packets are transferred in a manner
synchronous to the movement of the wafer by the positioning assembly.
Image data analysis may be conducted using conventional digital signal
processing methods.
With this interferometric detection apparatus, weak scattering from
particles in the range of 20 nm to 100 nm in size may be detected. This
compares to conventional scatterometry which can, at best, detect 80 nm or
larger particles. The amplitude of the scattered light is also dependent
on the refractive index of the particle. Since, in one embodiment, the
apparatus can provide spectroscopic data (which varies in response to
different wavelengths) of the deposit (particle), it can be used to
identify the composition of surface contaminants. Finally, since the
surface to be inspected moves in accordance with a predefined motion
relative to the scanning beam, erroneous detection of particles suspended
in the air may be avoided.
One advantage of the invention is that high resolution may be obtained. In
prior systems if two particles fell within one pixel no discrimination
could be obtained. The present invention allows this situation to be
distinguished and categorized by size and location.
These and other objects of this invention will become apparent to one
skilled in the art from the detailed description and the accompanying
drawings illustrating features of this invention by way of example.
BRIEF DESCRIPTION OF THE DRAWINGS
FIG. 1A is a diagrammatic representation of a linear array holographic
scatterometer embodiment according to the present invention.
FIG. 1B is an exaggerated, top view of a surface to be inspected.
FIG. 2 is a diagrammatic representation of an area array holographic
scatterometer embodiment according to the present invention.
FIG. 3 is a diagrammatic representation of the embodiment of FIG. 2
illustrating an alternate structure to attenuate a bias term, according to
the present invention.
FIG. 4 is a diagrammatic representation of a multi-channel holographic
scatterometer embodiment according to the present invention.
FIG. 5 is a diagrammatic representation of the embodiment of FIG. 4 using a
dual illumination path according to the present invention.
FIG. 6 is a diagrammatic representation of a holographic scatterometer
embodiment configured for particle composition analysis according to the
present invention.
FIG. 7 is a diagrammatic representation of a still yet another holographic
scatterometer embodiment according to the present invention.
FIG. 8 is a diagrammatic representation of an alternative embodiment to
that shown in FIG. 7, according to the present invention.
FIG. 9 is a diagrammatic representation of a preferred, alternate
arrangement of the optical assembly shown in FIGS. 1A, and 2-8.
FIG. 10 is a diagrammatic representation of another arrangement of the
optical assembly.
DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS
Referring now to the drawings wherein like reference numerals are used to
identify identical components in the various views, FIG. 1 shows an
apparatus 10 for inspecting a surface 12 of an object, such as a
semiconductor wafer 14 (preferably, unpatterned). Although the preferred
embodiments will be described in connection with the inspection of an
unpatterned semiconductor wafer, it should be understood that the present
invention is not so limited, and may be applied to other environments
where detecting (and alternately determining the composition of) particles
on an otherwise contaminant particle free surface is desirable or
required.
Before proceeding to a detailed description, a general overview of the
invention will be set forth. Small particles on an unpatterned wafer
surface are often undesirable and must be detected for remedial action. As
will be described in detail hereinafter, embodiments of the present
invention are configured to establish an interference fringe pattern when
such particles are present. In two dimensions, the fringe pattern is
similar to a series of concentric rings varying through a complete range
of intensities from substantial destructive interference (dark) to
substantial constructive interference (light). A section through the rings
appears similar to a sinusoid. Such a pattern may be detected in a fairly
straightforward manner if the fringe rings are "compressed". As used in
this specification, "compress" or "compression" means transforming a
dispersed interference signal (i.e., from a particle) and collapsing or
compressing it to a specific spatial location relative to a wafer surface.
This compression may be done in either one or two orthogonal dimensions
either optically or digitally (Note: the fringe pattern is a spatial two
dimensional phenomena as opposed to frequency which is a temporal or time
dependent determination). Once "compressed", the detection of the
interference pattern is representative of the presence of a foreign
particle on the wafer surface. Finally, since foreign particles in the air
do not move in synchronism with the wafer surface/CCD image movement, any
interference is effectively "smeared" out.
Referring now to FIG. 1, apparatus 10 includes a light source 16 configured
to generate a source light beam 18 and which further includes an expanding
and collimating assembly 20, an optical assembly 22, a photodetector 24, a
positioning assembly 26, and a controller 28.
Light source 16 may comprise conventional apparatus known to those of
ordinary skill in the art, including, but not limited to, Argon lasers
(multiple lines), diode-pumped solid-state lasers, quasi-coherent sources
such as Xenon arc lamps, and the like. For exemplary purposes only, an
embodiment of the disclosed invention may comprise a diode-pumped
solid-state (DPSS) laser having low noise, such as that commercially
available, referred to as a "Verdi" laser, manufactured by Coherent, Inc.,
having an output greater than 5.0 Watts, continuous, for example at a
single frequency such as 532 nanometers. In an alternate embodiment, a
multi-line light source may be used, such as that commercially available
and referred to as a "BeamLok 2080", available from Spectra Physics. Such
a multi-line light source may be a water-cooled ion-laser, configured for
25 watt, multi-line Argon or 5 watt multi-line Krypton operation (and
featuring active beam stabilization). Regarding wavelength selection, it
should be understood that, due to particle scattering effects, relatively
shorter wavelengths are preferable; however, this factor must be balanced
against the available illumination sources of sufficient intensity. The
fringe pattern does not change, substantially, for multiple wavelengths
(i.e., at least taken one at a time) or even for partially coherent
sources (although the modulation of the fringe pattern may be reduced).
Light source 16 further includes expanding and collimating assembly 20,
which is configured to produce an illumination beam, designated "I" in the
drawings, from source light beam 18, and for directing illumination beam I
toward, or in other words, in the direction of, surface 12. Assembly 20
may include a condenser lens, and a collimator, each of which may be
conventional apparatus known to those of ordinary skill in the art.
Illumination beam I, when incident upon a particle on the surface 12 of
object 14, creates a cone of backscattered light defining a scattered
beam, designated "S" in the drawings. Illumination beam I is also
specularly reflected by surface 12 to define a reference beam, designated
in the drawings as "R."
Optical assembly 22 is configured to generally direct the scattered light
beam S and the reference light beam R to photodetector 24. Optical
assembly 22, in the embodiment illustrated in FIG. 1A, includes a
beam-splitter 30, focusing optics such as a spherical lens 32 having an
optical axis 34, a focal point 35 and a focused image plane 36
respectively associated therewith, and a neutral density filter 38.
Beam-splitter 30, lens 32, and neutral density filter 38 may comprise
conventional apparatus known to those of ordinary skill in the art.
Photodetector 24 is provided for detecting an intensity (or light power) of
the light impinging thereon, and converting the same into a light
intensity signal, which may be a voltage signal whose amplitude
corresponds to light power. Photodetector 24 may comprise conventional
apparatus well-known to those of ordinary skill in the art. In the
embodiment illustrated in FIG. 1A, photodetector 24 comprises a camera
having a linear array charge-coupled device (CCD) 40, which has a
detection plane 42 associated therewith. For purposes of example only,
linear array CCD 40 may comprise commercially available components such as
model no. CL-CB-2048T, available from Dalsa, having 2048 pixels wherein
the intensity of each pixel is capable of being converted to a 12-bit
digital word, and further having a data rate of 20 megahertz. Such device
has a relatively good dynamic range, anti-blooming exposure control, and
multi-camera synchronization.
Positioning assembly 26 is configured to move wafer 14 such that surface 12
moves relative to the illumination beam I in accordance with a drive
control signal generated, for example, by controller 28.
FIG. 1B illustrates the orientation of the movement established by
positioning assembly 26 under the control of controller 28. FIG. 1B shows
wafer 14, including surface 12, disposed proximate positioning assembly 26
(shown in diagrammatic, fragmentary fashion). For purposes of the present
description, "scanning" occurs in a direction parallel to the x-axis. FIG.
1B also shows illumination beam I, in exaggerated form. In the illustrated
embodiment, the illumination beam remains fixed in absolute space, while
positioning assembly 26 moves wafer 14. Thus, for purposes of example
only, assume that it is desired to inspect the wafer surface 12 in an
increasingly positive x-axis direction (i.e., from left-to-right relative
to the orientation in FIG. 1B), then positioning assembly 26 would move
the wafer 14 in the negative x-axis direction (i.e., from right-to-left).
Of course, due to optical assembly 22 (and the inverting nature of a
spherical lens), the image of surface 12 moves in a negative x-axis
direction across photodetector 24. Expanding and collimating assembly 20,
and optical assembly 22 are, in one embodiment, operative to direct to
surface 12 an illumination beam I having an approximately diameter of 35
millimeters, and having an effective usable area in a center region
thereof approximately 20 millimeters on each side (i.e., a square region).
Moreover, in such an embodiment, the optics are such that the effective
pixel size (or sample area) at the wafer surface is approximately 5.0
microns. Apparatus 10, which employs a linear array CCD 40, images (and
captures the corresponding light intensity of) a "line" parallel to the
y-axis, shown in diagrammatic form in FIG. 1B as a dashed-line designated
"data."
After each scanning motion, which is preferably substantially continuous in
the x-axis direction, has been completed, positioning assembly 26 is
controlled by controller 28 to move or index the wafer 14 for the next
scan (i.e., based on the orientation in FIG. 1B, the indexing movement
occurs in the y-axis direction). After the entire surface 12 has been
inspected, and, provided that apparatus 10 is being used as an "in-line"
component in a fabrication process, positioning assembly 26 is further
controlled to move wafer 14 to the next processing station. Likewise,
positioning assembly, in such an "in-line" configuration, retrieves the
next wafer 14 to be inspected.
Positioning assembly 26 thus includes three distinct functions when
integrated into an "in-line" fabrication assembly line, namely: (i)
handling wafers to and from the inspection position; (ii) substantially
continuously moving the wafer in a scanning direction (i.e., x-axis
movement relative to the orientation of FIG. 1B); and (iii) indexing
movement for subsequent scans (i.e., y-axis movement based on the
orientation shown in FIG. 1B) Positioning assembly 26 may comprise
conventional apparatus known in the art for carrying out the
above-described functions. For example, in one embodiment, wafer
handling/positioning may be accomplished by a three-link arm "Gencobot IV"
robot, commercially available from Genmark, and which has a 24 inch reach,
and is capable of handling, in presently commercially available versions,
wafer sizes from 50 millimeters to 300 millimeters. Such a robot may be
used in combination with a standalone wafer positioner, such as also
available from Genmar | | |
