|
|
|
| United States Patent | 4652757 |
| Link to this page | http://www.wikipatents.com/4652757.html |
| Inventor(s) | Carver; Gary E. (Raritan Township, Hunterdon County, NJ) |
| Abstract | Individual defects in or near the surface of a silicon wafer (16) are
detected by directing a time-modulated laser beam (44), having an energy
level above the bandgap energy of the silicon material, towards the wafer.
The beam (44) is focused to a one to two micron spot (48) on the wafer
surface to photoexcite (i.e., pump) a high density of electrons and holes
which changes the infrared reflectance in the area of the pumped spot. A
probe beam (34) of infrared radiation is directed at the surface (0.126
square mm in area) of the substrate (16) and at a small angle thereto and
the reflection thereof monitored by a detector (54). The pumped spot (48)
is raster scanned within the area of the probe beam spot (38). The
detector (54) detects only that portion of the intensity of reflected
probe beam (34) that is modulated by the pump beam frequency to create a
video display having a high spatial resolution showing individual defects. |
|
|
|
Title Information  |
|
|
|
|
|
Drawing from US Patent 4652757 |
|
|
Method and apparatus for optically determining defects in a
semiconductor material |
|
|
|
|
|
| Publication Date |
March 24, 1987 |
|
|
|
|
|
| Filing Date |
August 2, 1985 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
Title Information  |
|
|
Claims  |
|
|
What is claimed is:
1. A method of detecting individual defects in or near the surface of a
semiconductor material, comprising the steps of:
directing an infrared beam of monochromatic light, having an energy level
below the bandgap energy of the semiconductor material, towards a first
area of the semiconductor surface;
simultaneously directing a time-modulated beam of monochromatic light,
having an energy level above the bandgap energy of the semiconductor
material, towards the surface of the semiconductor;
focusing the time-modulated light beam to a second area, substantially
smaller than, and within, the first area, the energy of the beam being
substantially absorbed by said surface to excite electrons and holes in or
near said material surface to alter the reflectivity of said surface;
detecting the intensity of the infrared light beam reflected from the
semiconductor surface; and
processing only that portion of the detected infrared beam intensity that
is modulated at the frequency of the time-modulated light beam to locate
defects in or near the semiconductor surface.
2. The method as set forth in claim 1, comprising the steps of:
raster scanning the time modulated light beam to move the second area
within the first area; and
detecting the intensity of the light from only the infrared beam reflected
from the surface at the modulated frequency.
3. The method as set forth in claim 1, wherein:
the first area is approximately 0.126 square mm; and
the second area is a spot of approximately 1 to 2 microns in diameter.
4. The method as set forth in claim 1, wherein:
the semiconductor material is silicon.
5. The method as set forth in claim 1, wherein:
the semiconductor material is InP.
6. The method as set forth in claim 1, wherein:
the semiconductor material is InGaAsP.
7. The method as set forth in claim 1, wherein;
the infrared beam has an angle of incidence to the semiconductor surface of
79.degree. to a plane normal to said surface.
8. The method as set forth in claim 1, wherein:
the time-modulated beam was generated by an argon laser; and
the infrared beam was generated by a Co.sub.2 laser. |
|
|
|
|
Claims  |
|
|
Description  |
|
|
TECHNICAL FIELD
The invention relates to testing the quality of semiconductor wafers. In
particular, wafer defects are determined using non-destructive techniques.
BACKGROUND OF THE INVENTION
In the fabrication of semiconductor devices, the surface of the
semiconductor material in which the devices are fabricated must be
substantially free of both physical and crystalline defects. A high degree
of crystalline perfection is necessary to produce reliable devices having
good electrical properties. In order to control the properties of such
devices, it is necessary to be able to determine the quality of
semiconductor material that is being used to make the devices.
Defects in semiconductor material (e.g., silicon, InP, InGaAsP or the like)
include dislocations, stacking faults, oxygen precipitates and heavy metal
precipitates. Such defects, which are often one to two microns in size,
have been detected using a variety of techniques including etch pit
analysis, x-ray topography and electron beam induced current.
Etch pit analysis involves etching a silicon wafer and then viewing the
pits that grow in the defective areas under an interference microscope.
Individual stacking faults, dislocations and saucer pits can be
distinguished after etching by the shape of the pits that each defect
causes to grow. X-ray topography is accomplished by setting up an x-ray
beam, the wafer and a film to catch Bragg-reflected rays. The relative
angular positions of these three components are such that the Bragg law is
satisfied when the beam strikes good material, but is not satisfied when
it encounters defects. The film and sample are translated such that the
reflected beam intensity for all points on the wafer are mapped onto the
film. Spatial resolutions of 1-10 microns can be attained after 3 to 20
hours of exposure.
Electron beam induced current (EBIC) is accomplished inside a scanning
electron microscope (SEM). The SEM beam is used to induce carriers within
a sample, which are then separated by an electric field. The field is
created by either a pn junction within the sample, a Schottky junction on
the surface of the sample, or by mounting the sample within an external
field. Contact to the sample is generally accomplished with silver paste
or spring loaded fine wires. As the SEM beam is rastered in x-y plane the
collected carriers provide intensity modulation of an x-y display. Carrier
recombination at defects allows them to appear dark on the display screen,
down to a spatial resolution of 0.5 micron. Although these techniques
image defects with high spatial resolution in semiconductor material, they
are either destructive, time consuming and/or require a vacuum.
One non-destructive optical technique for determining electrical
non-uniformities in semiconductor wafers is described in U.S. Pat. No.
4,211,488 to Kleinknecht which issued on July 8, 1980. That patent makes
use of the fact that crystalline imperfections or doping striations in a
semiconductor wafer cause lower carrier lifetime and/or mobility during
photoexcitation and therefore change the infrared reflectance of the
material. The electrical non-uniformities or defective areas are detected
by irradiating an area of the semiconductor wafer with a beam of
monochromatic light having energy greater than the bandgap energy of the
semiconductor wafer material. This will photoexcite (i.e., pump) a high
density of electrons and holes which changes the infrared reflectance at
the pumped area. The same surface area of the wafer is simultaneously
irradiated with a second beam of monochromatic light having an energy less
than the bandgap of the semiconductor material, whereby part of the second
beam is reflected from the surface.
If the monitored area has moderate to low defect density and high carrier
mobility, the reflectance of the surface will change during
photoexcitation and the intensity of the reflected second beam will also
change. However, if there is a high defect density within the area the
reflectance of the surface will not change during photoexcitation and the
intensity of the reflected second beam will remain unchanged. The
intensity of the reflected beam is detected and the magnitude thereof is a
measure of the carrier mobility and recombination time which is directly
related to the density of the surface or near surface defects in the
semiconductor material. The light beams in the Kleinknecht patent
simultaneously illuminate an area of about 0.25 square mm.
Although such a technique can effectively provide information as to the
average carrier lifetime and mobility over the 0.25 square mm area, it
cannot resolve individual defects of one to two microns in size. There are
two fundamental reasons for this fact. First, the laser providing the
below-bandgap energy emits long wavelength light in the infrared part of
the spectrum. Since basic diffraction theory predicts that minimum
obtainable spot size is proportional to the f number times the wavelength,
infrared light having wavelengths of interest for defect detection can be
focused to spots no smaller than 10 to 20 microns. Second, the probe beam
in the Kleinknecht patent has a high angle of incidence with respect to a
normal to the wafer surface. This high angle leads to a further
enlargement of the probe beam spot. Therefore, such a technique cannot
focus the infrared beam to a small enough spot to resolve individual
defects, of micron size, due to accepted basic optical theory. However,
there is clearly great interest in resolving these micron-sized defects
due to their influence on VLSI circuits having micron-sized features.
Accordingly, there is a need for a non-destructive defect detection system
in which individual defects of one to two microns in size can be resolved.
SUMMARY OF THE INVENTION
The foregoing need is met with the instant method of detecting individual
defects in or near the surface of a semiconductor, comprising: directing
an infrared beam of monochromatic light, having an energy level below the
bandgap energy of the semiconductor material, towards a first area of the
semiconductor surface; simultaneously directing a time-modulated beam of
monochromatic light, having an energy level above the bandgap energy of
the semiconductor material, towards the surface of the semiconductor
surface; focusing the time-modulated light beam to a second area,
substantially smaller than, and within, the first area, the energy of the
beam being substantially absorbed by said surface to excite electrons and
holes in or near said material surface; detecting the intensity of the
infrared light beam reflected from the semiconductor surface; and
processing only that portion of the detected infrared beam intensity that
is modulated at the frequency of the time-modulated light beam to locate
individual defects in or near the semiconductor surface.
BRIEF DESCRIPTION OF THE DRAWINGS
FIG. 1 is a schematic diagram of a known defect detection system;
FIG. 2 is a schematic diagram of the instant defect detection system; and
FIG. 3 is a block diagram of the instant defect detection system.
DETAILED DESCRIPTION
FIG. 1 schematically depicts the known technique described in the
aforementioned Kleinknecht patent which is incorporated by reference
herein. A laser 12 modulated at 10 Hz and a laser 14 both direct light
beams at a semiconductor substrate 16. The laser 12 outputs a pump beam
that provides energy which is above the bandgap of the material of the
substrate 16 while the laser 14 emits a beam that provides energy which is
below the bandgap of the material of the substrate. The impingement on the
substrate 16 of both laser beams is coincident in an area 17 which is
approximately 0.25 square mm. If there are moderate to low defect
densities in or near the surface of the substrate 16, the infrared
reflectance of the surface changes and therefore the amount of reflected
light from the laser 14 changes. However, if there are large defect
densities in or near the substrate surface the reflectivity thereof is
substantially unchanged during photoexcitation and the intensity of the
reflected beam 14 will be substantially unchanged.
The beam from the laser 14 is reflected from the surface of the substrate
16 and impinges upon a detector 18 which measures its intensity. All of
the intensity information is converted into electrical impulses which are
forwarded to an oscilliscope (not shown) where the strength of the
electrical signal is displayed in volts. As hereinbefore indicated, such a
technique provides information as to the average carrier lifetime and
mobility over a 0.25 square mm area but it cannot resolve individual
defects of one to two microns in size.
The instant technique, depicted by the schematic diagram in FIG. 2,
overcomes the foregoing problem. A below-bandgap laser 32 directs a
monochromatic probe beam 34 of infrared light (10.6 .mu.m) towards a
silicon substrate 16 and incident thereto at the Brewster angle. The probe
beam 34 passes through a lens 36 which focuses the beam on the substrate
16. When the probe beam 34 has an incidence angle of 79.degree. to a plane
normal to the substrate 16 a probe spot 38 of approximately 0.126 square
mm is formed on the substrate.
An above-bandgap laser 42 simultaneously directs a pump beam 44 towards the
substrate 16. A lens 46 focuses the beam 44 to a small pump spot 48 of
about 2 .mu.m in diameter on the substrate 16. The focused pump beam 44
will alter the infrared reflectivity of the semiconductor material as
hereinbefore described. The small pump spot 48 falls within the much
larger probe spot 38. The relative size of the pump spot 48 and the probe
spot 38 are not to scale in FIG. 2 but are shown in this manner for
purposes of clarity of the drawings. Additionally, the pump beam 44 passes
through an acoustic cell (not shown) to modulate the beam at 30 kHz in an
exemplary embodiment.
The probe beam 34 is reflected from the surface of the substrate 16 and is
relayed by a lens 52 for impingement upon a detector 54 where the
reflected light intensity is converted to proportional electrical signals.
The detector 54 and associated electronics process only those signals
resulting from the reflected light from the probe beam 34 that is
modulated at the 30 kHz frequency of the pump beam 44. As hereinbefore
indicated the pump beam 44 will change the infrared reflectivity of the
surface of the substrate 16 in the absence of defects but the infrared
reflectivity is left substantially unchanged when the pump spot falls on a
defect. Accordingly, that portion of the probe beam 34 impinging on the
pump spot 48 will be modulated at the 30 kHz frequency of the pump beam 44
which alters the surface reflectivity at that frequency.
Additionally, the modulated pump beam 44 may be raster scanned within the
0.126 square mm spot of the probe beam 34 by selectively moving the pump
beam and/or the substrate 16 in a well known manner. Once the probe spot
38 is fully scanned by the pump spot 48, it may then be moved to another
location on the wafer 16 and the pump beam spot 48 again raster scanned
therein. This process can be repeated until the full surface of the wafer
16 has been scanned and the reflections detected and forwarded to a video
apparatus (not shown) for visual display.
FIG. 3 is a block diagram of exemplary optics and electronics of the
instant defect detection system 50. A 300 mw water cooled argon pumping
laser 52, having a wavelength of 0.488 .mu.m, is directed towards the
silicon substrate 16 via a beamsplitter 54, an acoustic optical (AO)
modulator 56 which time modulates the beam 44 at 30 kHz, a galvanometer
mirror 58 and focusing optics 62. About 30 mw of light power from the
argon pumping laser 52 actually impinges on the surface of the wafer 16,
which is a sub-damage intensity density. About 1.5 watts from a five-watt,
air-cooled CO.sub.2 probe laser 64 is directed to the same location on the
substrate 16 by reflecting the beam 34 from a fixed mirror 66 and through
a focusing lens 36 to form an angle of incidence of 79.degree. with a
plane normal to the substrate. The probe beam 34 is reflected from the
surface of the wafer 16 and further reflected from a fixed mirror 68 onto
a HgCdTe detector 72 which has an output connected to an input of a lockin
amplifier 74. The detector 72 and the lockin amplifier 74 process only
that portion of the intensity of the reflected probe beam 34 that is
modulated at the 30 kHz frequency. Advantageously, such a technique
transforms the spot size typical for visible radiation of the pump beam 44
into the longer infrared radiation of the probe beam 34.
A light detector 76 monitors the position of the pump laser beam 44 by
detecting reflections from the beamsplitter 54 and forwarding that
information to the input of the lockin amplifier 74. A signal from the
drive connection of the lockin amplifier 74 is forwarded to the acoustic
optical modulator 56 to modulate the pump beam 44 at the same frequency as
that at which the inputs of the lockin amplifier 74 are adjusted to
receive. The output of the lockin amplifier 74 is sent to a computer 77
having outputs to a scan controller 78 and a video display 82. The output
of the scan controller 78 is connected to a movable table 80 and the
galvanometer controlled mirror 58 to control the relative positions of the
table and the mirror 58.
The substrate 16 is mounted on a quartz vacuum chuck (not shown) which is
supported on the movable table 80 which is capable of movement in the X,
Y, Z and .theta. directions. The Z stage is used to bring the substrate 16
into focus. The focused position in the Z direction may change from one
scan location to another due to the depth of focus of the lens used. The Y
and .theta. stages are used to access various points on the surface of the
wafer 16. The X stage and rotatable galvanometer mirror 58 are used to
generate the raster scan in a well known fashion at approximately a thirty
second frame rate. All four stages and the galvanometer mirror 58 operate
under the control of the computer 77.
It is to be understood that the embodiments described herein are merely
illustrative of the principles of the invention. Various modifications may
be made thereto by persons skilled in the art which may embody the
principles of the invention and fall within the spirit and scope thereof.
* * * * *
|
|
|
|
|
Description  |
|