 |
Description  |
|
|
BACKGROUND OF THE INVENTION
The present invention relates generally to a method and system for
inspecting a polycrystalline semiconductor film. More specifically, the
invention relates to an inspecting method and system for measuring grain
sizes of semiconductor grains forming a polycrystalline semiconductor
film.
In general, a thin film transistor (TFT) using a polycrystalline silicon
(poly-Si) has an advantage in that the crystal has a mobility 10 through
100 times as high as that of a TFT using an amorphous silicon (a-Si).
Therefore, there has been studied and developed a drive circuit integrated
thin film transistor-liquid crystal display (TFT-LCD) wherein a thin film
transistor of a polycrystalline silicon is not only used as a pixel
switching element for the liquid crystal display (LCD), but it is also
used as a peripheral drive circuit for the liquid crystal display, to form
the thin film transistors for the pixel and the drive circuit on the same
substrate. In such a drive circuit integrated TFT-LCD, there is a
correlation that the mobility of crystals of a thin film transistor of a
polycrystalline silicon increases as the crystal grain size increases, as
can be seen from a characteristic diagram of FIG. 1 which shows the
relationship between the crystal grain size and mobility. For that reason,
there is an important problem of how to measure the crystal grain size.
Conventionally, the crystal grain size is measured by a scanning electron
beam microscope (SEM) or the like after a grain boundary is selectively
removed by an etching, such as SECO etching, or a cross section taken
along a thickness direction of a substrate is observed by a transmission
electron beam microscope (TEM). However, in such conventional inspection
methods, it takes at least 2 hours to observe the grain size. In order to
shorten the measuring time, it has been proposed to use an atomic beam
frequency microscope (AFM). Although the grain size can be observed and
measured by the AFM, it takes about 30 minutes to observe a one point to
analyze the grain size.
In addition, there is well known a basic method for observing a grain size
by means of an optical microscope having a magnification of 500 thorough
1000 by using the variation in irregularity on the surface of a film as an
index of the grain size. However, since this method is greatly relied on
the human naked eye, the measured results are easily influenced by the
difference among individuals, so that there are problems in that quite
accurate results can not be obtained and the measured results are not
quantitative.
Moreover, there may be considered a grain size analysis using an
ellipsometer which can measure a grain size in a non-destructive and
non-contact manner in a short measuring time of 5 seconds per one point
This conventional measuring method using the ellipsometer is used for
analyzing an object having a flat surface, such as a silicon oxide film on
a single crystal. However, if this method is used for analyzing a
polycrystalline silicon, there is a problem in that it is difficult to
construct an analysis model, e.g., it is difficult to quantify the grain
size, mobility and so forth. Particularly in the case of the conventional
ellipsometer measuring method, a polycrystalline silicon produced by the
excimer laser annealing (ELA) method has irregularities on the surface
even if the thickness thereof is, e.g., about 50 nm, so that it is
particularly difficult to quantify the grain size, mobility and so forth.
On the other hand, in a polycrystalline semiconductor film inspecting
method according to the present invention, a polycrystalline semiconductor
film having surface irregularities produced by the ELA is used as a
sample, and a polycrystalline semiconductor film is also used as a
reference sample.
For example, when 12-inch and 15-inch or more drive circuit integrated thin
film transistor-liquid crystal displays are produced, the mean grain sizes
of the polycrystalline silicon are suitably 0.25 .mu.m and 0.45 .mu.m or
more, respectively. Because, in the case of a display size of 12 inches,
the mobility of crystals in n-channel lightly doped drain-thin film
transistors (n-ch LDD-TFT) having a grain size of 0.25 .mu.m or less is
100 cm.sup.2 /Vs or less, so that it is difficult to drive a 12-inch class
LCD. In addition, because, in the case of a display size of 15 inches, the
mobility of crystals in n-channel lightly doped drain-thin film
transistors (n-ch LDD-TFT) having a grain size of 0.45 .mu.m or less is
120 cm.sup.2 /Vs or less, so that it is difficult to drive a 15-inch class
or more LCD. Therefore, the polycrystalline silicon films of the drive
circuit integrated thin film transistor-liquid crystal displays are
preferably prepared so as to have grain sizes of about 0.25 .mu.m and
about 0.45 .mu.m, respectively. However, it is impossible to accurately
measure the polycrystalline silicon films having this range of grain size
in a short time.
In conventional methods, e.g., in a method utilizing the relationship
between mean grain sizes of a polycrystalline silicon and proportions of
compositions of the polycrystalline silicon when the polycrystalline
silicon is represented as a mixture of an amorphous silicon and a
crystalline silicon (c-Si), as shown in FIG. 2, or in a method utilizing
the relationship between mean grain sizes of a polycrystalline silicon and
proportions of compositions of the polycrystalline silicon when the
polycrystalline silicon is represented as a mixture of an amorphous
silicon, a polycrystalline silicon and a crystalline silicon, as shown in
FIG. 3, there is a problem in that there is no repeatability in a sample
having a thickness difference of .+-.5%, so that it is not possible to
achieve accuracy which can be utilzed in the measurement of grain sizes.
Because both of the thickness and quality of a film are simultaneously
calculated as parameters during analysis. That is, if only the thickness
of the film is determined before the calculation of the quality of the
film, the quality of the film is different from that of an actual sample.
Moreover, if the quality of the film is calculated using the determined
thickness of the film, it is required to calculate the thickness again
using the calculated quality of the film.
In addition, the surface of polycrystalline silicon is coated with a
natural oxide film, and a polycrystalline silicon film formed by the ELA
method has surface irregularities as an inherent property, so that there
is a problem in that conventional analyzing methods for use in
polycrystalline silicon film having a flat surface can not be used. An
example of such conventional polycrystalline silicon inspecting methods is
disclosed in 1992 American Institute of Physics, J. Appl. Phys. 72(8),
Oct. 15, 1992, "Comparative study of thin poly-Si films grown by ion
implantation and annealing with spectroscopic ellipsometry, raman
spectroscopy, and electron microscopy". It is reported in this literature
that the peak width at half height obtained by differentiating, two times,
the peak near 4 eV appeared with a dielectric constant (.di-elect cons.)
of a polycrystalline silicon is used as a parameter.
However, according to the above described conventional inspecting method,
there is a problem in that the grain size of a polycrystalline silicon of
0.8 .mu.m is estimated as 0.05 .mu.m. In addition, an object to be
estimated must have a grain size of 0.15 .mu.m or less, and this has a low
mobility, so that there is a problem in that this can not be used for
forming a TFT channel for a liquid crystal display.
As described above, in any conventional inspection methods, it is difficult
to quantify the grain size, mobility and so forth of crystals in analysis
of the grain size, and it is difficult to accurately measure the crystal
grain size in a short time.
SUMMARY OF THE INVENTION
It is therefore an object of the present invention to eliminate the
aforementioned problems and to provide a method and system for inspecting
a polycrystalline semiconductor film, which can accurately measure the
grain size of the polycrystalline semiconductor film in a short time in a
non-destructive and non-contact manner.
In order to accomplish the aforementioned and other objects, according to
one aspect of the present invention, a polycrystalline semiconductor film
inspecting method comprises the steps of: calculating dependencies on
wavelength of refractive index and damping coefficients of a plurality of
standard samples including at least a polycrystalline semiconductor film;
calculating dependency on wavelength of a refractive index and a damping
coefficient, and a thickness, of an estimated sample consisting of a
polycrystalline semiconductor film; comparing the dependencies on
wavelength of the refractive index and the damping coefficient of the
estimated sample, with those of the standard samples so as to derive the
compared results as indexes; and deriving a correlation between the
thickness of the estimated sample and indexes derived as the comparison
results. Thus, the standard sample and the sample to be estimated, which
serves as the object to be inspected, are measured, so that it is possible
to quantify the optically measured results and it is possible to
accurately measure a crystal grain size in a short time even in a
non-destructive and non-contact manner.
The dependencies on wavelength of the refractive index and the damping
coefficient of the polycrystalline semiconductor film serving as the
object to be estimated may be compared with those of at least one of an
amorphous semiconductor and a crystalline semiconductor. Thus, it is
possible to more accurately measure the grain size.
In addition, a correlation between the thickness of the polycrystalline
semiconductor film and the indexes serving as the comparison results may
be derived, and the polycrystalline semiconductor film serving as the
object to be estimated may be annealed while adjusting energy in
accordance with the correlation. Thus, the polycrystalline semiconductor
film can have a desired grain size.
Moreover, .PSI. and .DELTA., which represent a ratio of a reflectance of a
p-polarized light to that of an s-polarized light by
tan(.PSI.).multidot.exp(i.DELTA.), may be substituted for the dependencies
on wavelength of the refractive index and the damping coefficient.
According to another aspect of the present invention, a polycrystalline
semiconductor film inspecting method comprises the steps of: irradiating a
polycrystalline semiconductor film, which is formed on a substrate, with a
light, and detecting dependence on wavelength of the intensity of a
reflected light; and comparing the dependence on wavelength of the
intensity of the reflected light with a sample data, and calculating a
crystal grain size of the polycrystalline semiconductor film or data
correlate therewith.
Furthermore, the functional formulae for deriving dependencies on
wavelength should not be limited to specific formulae, but any functional
formulae may be used as long as the formulae meet a predetermined
correlation. That is, the feature of the present invention is that data
are previously derived on the basis of the measured results of a plurality
of standard samples, and the same data on a sample to be estimated are
measured to derive the inspected results, or a polycrystalline
semiconductor film meeting desired conditions is derived by processing the
derived results while being compared with the results of the standard
samples.
BRIEF DESCRIPTION OF THE DRAWINGS
In the drawings:
FIG. 1 is a graph showing the relationship between mobility of a thin film
transistor of a polycrystalline silicon and mean grain sizes of the
polycrystalline silicon;
FIG. 2 is a graph showing the relationship between mean grain sizes of a
polycrystalline silicon and proportions of compositions of the
polycrystalline silicon when the polycrystalline silicon is represented as
a mixture of an amorphous silicon and a crystalline silicon;
FIG. 3 is a graph showing the relationship between mean grain sizes of a
polycrystalline silicon and proportions of compositions of the
polycrystalline silicon when the polycrystalline silicon is represented as
a mixture of an amorphous silicon, a polycrystalline silicon and a
crystalline silicon;
FIG. 4 is a sectional view showing the concept of the first preferred
embodiment of a polycrystalline semiconductor film inspecting method
according to the present invention;
FIG. 5 is a flow chart showing steps of the first preferred embodiment of a
polycrystalline semiconductor film inspecting method according to the
present invention;
FIG. 6 is a graph showing the relationship between thickness of a
polycrystalline silicon film and silicon crystal grain sizes in the first
preferred embodiment;
FIG. 7 is a sectional view of an embodiment of a thin film transistor;
FIG. 8 is a conceptual drawing of a preferred embodiment of a
polycrystalline silicon film inspecting system according to the present
invention;
FIG. 9 is a block diagram of the second preferred embodiment of a
polycrystalline semiconductor film inspecting system according to the
present invention;
FIG. 10 is a graph showing the relationship between dependence on
wavelength of a refractive index of a laminated structure of a glass
substrate, a silicon nitride film, a silicon oxide film and a
polycrystalline silicon film, and measuring energy to a silicon oxide
film;
FIG. 11 is a graph showing the relationship between dependence on
wavelength of a refractive index of a laminated structure of a glass
substrate, a silicon nitride film, a silicon oxide film and a
polycrystalline silicon film, and dependence on thickness of a silicon
nitride film;
FIG. 12 is a graph showing the relationship between proportions of
compositions of a polycrystalline silicon (p-Si) film having a thickness
of 0.3 .mu.m and mean grain sizes of the polycrystalline silicon film when
the polycrystalline silicon film is represented by a polycrystalline
silicon film having a thickness of 0.3 .mu.m and a polycrystalline silicon
film having a thickness of 0.5 .mu.m;
FIG. 13 is a graph showing the relationship between proportions of
compositions of a polycrystalline silicon film having a mean grain size of
0.3 .mu.m and film pressures of the polycrystalline silicon film;
FIG. 14 is a graph showing the relationship between dependence of a grain
size and the maximum refractive index of polycrystalline silicon films
having mean grain sizes of 0.1 .mu.m, 0.3 .mu.m and 0.5 .mu.m;
FIG. 15 is a graph showing dependence on wavelength of a refractive index
in the third preferred embodiment of a polycrystalline semiconductor film
inspecting method according to the present invention;
FIG. 16 is a graph showing dependence on wavelength of a dielectric
constant (.di-elect cons.) of a polycrystalline silicon in the third
preferred embodiment of a polycrystalline semiconductor film inspecting
method according to the present invention; and
FIG. 17 is a graph showing the relationship between mean grain sizes of
polycrystalline silicon and peak widths at half height in the third
preferred embodiment.
DESCRIPTION OF THE PREFERRED EMBODIMENTS
Referring now to the accompanying drawings, the preferred embodiments of a
method and system for inspecting a polycrystalline semiconductor film
according to the present invention will be described in detail below.
First, the concept of a polycrystalline semiconductor used in the
specification will be described. In semiconductors for use in TFT-LCDs or
the like, it is well known that impurities, such as boron (B), phosphorus
(P) and arsenic (As), are implanted into the semiconductors in order to
adjust the threshold voltage Vth of TFTs. Therefore, throughout the
specification, the semiconductor films for use in the present invention
include thin films formed by implanting impurities, such as boron (B),
phosphorus (P) and arsenic (As), into semiconductors, such as silicon
(Si), germanium (Ge), gallium arsenide (GaAs) and cadmium selenide (CdSe).
First, referring to FIGS. 4 through 6, the first preferred embodiment of a
polycrystalline semiconductor film inspecting method serving as a basic
principle according to the present invention will be described below. In
the first preferred embodiment of a polycrystalline semiconductor film
inspecting method according to the present invention, first, second and
third samples serving as standard samples are separately formed as shown
in FIG. 4. While three types of samples, i.e., the first through samples,
serving as standard samples have been formed in this preferred embodiment,
only two types of polycrystalline semiconductor film samples, i.e., only
the first and second samples, as the minimum number of samples may be
formed to apply the present invention.
In the first preferred embodiment shown in FIG. 4, the first sample is a
polycrystalline silicon film having a crystal grain size of 0.3 .mu.m
formed on a glass substrate. The second sample is a polycrystalline
silicon film having a crystal grain size of 0.5 .mu.m formed on a glass
substrate, and the third sample is an amorphous silicon film formed on a
glass substrate. These standard samples are previously prepared, and
dependence on wavelength of a light refractive index of each of the
standard samples and dependence on wavelength of a damping coefficient
thereof are previously measured to calculate the values thereof.
Then, as shown on right side of FIG. 4, a polycrystalline silicon film
serving as a sample to be estimated, which is an object to be inspected,
is prepared, and the sample to be estimated is irradiated with light to
measure a light refractive index and a light damping coefficient of the
polycrystalline silicon film to calculate dependence on wavelength of the
refractive index and dependence on wavelength of the damping coefficient.
Referring to FIG. 5, processing steps of the first preferred embodiment of
a polycrystalline semiconductor film inspecting method according to the
present invention will be described. First, at step ST1, dependence on
wavelength of a refractive index and dependence on wavelength of a damping
coefficient for each of the first through third polycrystalline silicon
film samples serving as the standard samples are calculated. Then, at a
second calculating step ST2, dependence on wavelength of a refractive
index, dependence on wavelength of a damping coefficient, and a thickness
of a polycrystalline semiconductor film serving as an object to be
estimated are calculated. Then, at a comparison step ST3, the dependencies
on wavelength of the refractive index and damping coefficient for the
polycrystalline semiconductor film serving as the object to be estimated
are compared with those for the polycrystalline semiconductor films
serving as the standard samples, and the results thereof are obtained as
indexes. Then, a correlation step ST4, the correlation between the
thickness of the polycrystalline semiconductor film and the indexes
obtained at the comparison step is derived. Finally, in accordance with
the correlation obtained at the correlation step SP4, anneal is carried
out while adjusting energy of an energy beam irradiated on the
polycrystalline silicon film serving as the sample to be estimated.
Thus, in view of the previously obtained dependencies on wavelength of the
refractive index and damping coefficient of the standard samples, the
dependence on wavelength of the polycrystalline silicon film is measured,
and a beam is implanted while adjusting the dose of energy while it is
compared with the values obtained using the standard samples, so that it
is possible to obtain a desired polycrystalline silicon film. The
relationship between the crystal grain sizes of polycrystalline silicon
and the thickness of the silicon film in that case is shown in FIG. 6. As
can be seen from this graph, there is a certain correlation between the
thickness of the polycrystalline silicon film and the crystal grain size,
so that it is possible to obtain a desired polycrystalline film by
adjusting the crystal grain size in accordance with the thickness of the
film.
The second preferred embodiment of a polycrystalline semiconductor film
inspecting method according to the present invention will be described
below.
First, referring to FIG. 7, a thin film transistor used for, e.g., 12-inch
class drive circuit integrated liquid crystal display, will be described.
As shown in FIG. 7, a buffer layer 3 having a laminated structure of a
silicon nitride (SiNx) and silicon oxide (SiOx) is formed on an insulating
substrate 2. On the buffer layer 3, there are formed a channel 4 of a
polycrystalline silicon wherein a peak value of a refractive index (n) at
2.95.about.3.55 eV is limited to 6.28 or less, e.g., of a polycrystalline
silicon serving as a polycrystalline semiconductor having a mean grain
size of 0.24 .mu.m.about.0.45 .mu.m, and a drain region 5 and a source
region 6 which are formed by impurity implantation.
Moreover, a gate insulator film 7 is formed on the buffer layer 3 including
the channel 4, the drain region 5 and the source region 6. A gate
electrode 8 is formed on the gate insulator film 7 above the channel 4. An
interlayer insulator film 9 is formed on the gate electrode 8. The
interlayer insulator film 9 has contact holes 10 and 11, which are filled
with a drain electrode 12 and a source electrode 14, respectively, which
are connected to the drain region 5 and the source region 6, respectively.
Furthermore, in order for the 12-inch drive class circuit integrated liquid
crystal display to achieve a good display, the mean grain size of the
polycrystalline silicon must be in the range of from 0.24 .mu.m to 0.45
.mu.m, so that the peak value appeared at 3.25.+-.0.30 eV in the
polycrystalline silicon formed by the excimer laser anneal (ELA) must be
set to be less than 6.28 eV.
While the thin film transistor shown in FIG. 7 has been described as an
example of an coplanar type transistor, the present invention should not
be limited thereto, but the invention may be applied to a stagger type or
reverse stagger type thin film transistor.
Referring to FIG. 8, the second preferred embodiment of a system for
inspecting a thin film transistor of a polycrystalline silicon, according
to the present invention, will be described below. In FIG. 8, a
polycrystalline silicon thin film inspection system 21 comprises an XeCl
excimer laser anneal system 22, and a spectral ellipsometer system 23
integral therewith.
The excimer laser anneal system 22 has an excimer laser oscillator 25
having a variable fluent which is one of forming conditions. After a laser
beam from the excimer laser oscillator 25 reflects on an optical mirror
26, the waveform of the laser beam is adjusted by a beam homogenizer 27,
and the waveform adjusted laser beam is supplied to a casing 28 of the
spectral ellipsometer system 23. The casing 28 houses therein: a position
adjustable, movable table 29 for mounting thereon a sample 30; a light
source 31 for irradiating on the sample 30 with a light; a chopper 32;
polarizers 33 and 34; and a rotating analyzer 35. The amorphous silicon
sample 30 mounted on the table 29 is also irradiated with the light which
has reflected on the optical mirror 26.
The spectral ellipsometer system 23 has the light source 31 for irradiating
the sample 30 with light. The sample 30 is irradiated with the light from
the light source 31 via the chopper 32 and the polarizer 33. Then, the
light reflects on the sample 30 to be incident on the rotating analyzer
35. Since the light source 31 and the rotating analyzer 35 are connected
to a personal computer 36, the outgoing light from the light source 31 and
the light received by the rotating analyzer 35 are analyzed by means of
the personal computer 36. The results are fed back to the excimer laser
oscillator 25.
At this time, if the grain size of the sample 30 is too large, the fluent
of the laser beam from the excimer laser oscillator 25 is reduced, and if
the grain size of the sample 30 is too small, the fluent of the laser beam
is increased. The feedback of the fluent may be carried out in a sheet of
the sample 30. In this case, the fluent may be changed while annealing the
sample 30. Alternatively, the thickness of an amorphous silicon film may
be measured immediately before the excimer laser anneal, to be used as one
of setting information on the fluent of the excimer laser oscillator 25.
FIG. 9 shows the more detailed construction of the second preferred
embodiment of a polycrystalline silicon film inspecting system according
to the present invention. As shown in FIG. 9, the polycrystalline silicon
film inspecting system comprises an excimer laser oscillator system 22, a
spectral ellipsometer system 23 and inspection control system 40. The
inspection control system 40 comprises standard sample calculating means
41, estimated sample calculating means 45, comparing means 46 for
comparing the values calculated by the calculating means 41 and 45,
correlation means 47 for deriving a correlation between both samples on
the basis of the output of the comparing means 46, and adjusted value
outputting means 48 for adjusting the quantity of a laser beam, which is
oscillated by the excimer laser oscillator system 22, on the basis of the
output of the correlation means 47. Furthermore, also in the second
preferred embodiment similar to the first preferred embodiment, the
standard sample calculating means 41 has first through third sample
calculating parts 42 through 44, and is designed to previously calculate
dependence on wavelength of a refractive index and dependence on
wavelength of a damping coefficient for the first through third samples as
shown in FIG. 4.
An inspection method using the polycrystalline silicon thin film inspecting
system 21 shown in FIG. 8 will be described below. For example, the sample
30 is a polycrystalline silicon film having a laminated structure of a
silicon nitride (SiNx) film having a thickness of 50 nm, a silicon oxide
(SiOx) film having a thickness of about 100 nm and an amorphous silicon
film of about 55 nm, which is formed on a glass substrate by the excimer
laser anneal method using a XeCl excimer laser. For example, laser beams
having irradiation energies of about 341 mJ/cm.sup.2 and 305 mJ/cm.sup.2
are irradiated 26 times, respectively. In addition, it is assumed that the
mean grain sizes of the polycrystalline silicon film are about 0.52 .mu.m
and about 0.31 .mu.m, and the surface irregularities are about 22 nm and
about 53 nm, respectively.
In addition, a sample having a laminated structure of a silicon nitride
(SiNx) film having a thickness of about 50 nm, a silicon oxide (SiOx) film
having a thickness of about 100 nm and an amorphous silicon having a
thickness of about 55 nm on the glass substrate is prepared as a standard
sample.
Then, the polarization characteristics of these standard samples are
measured by means of a spectral ellipsometer. Then, the polarization
characteristics are analyzed to calculate dependencies on wavelength of
refractive indexes (n) and damping coefficients (k) of the amorphous and
polycrystalline silicon samples. The dependencies on wavelength of
refractive indexes (n) of the obtained standard samples are represented by
functions, n.sub.a-Si (.lambda.) and n6(.lambda.).
Then, the polarization characteristics of the polycrystalline silicon
sample, the grain size of which is to be measured, i.e., the
polycrystalline silicon sample comprising a silicon nitride film having a
thickness of about 50 nm, a silicon oxide film having a thickness of about
100 nm and a polycrystalline silicon film having a thickness of about 55
nm, which have been laminated on the glass substrate, and which serves as
a sample to be inspected, are measured by a spectral ellipsometer.
In this measurement, the measuring energy is preferably in the range of
from 2.0 eV to 5.0 eV. Because the range of measuring energy has an
influence on variations in thickness of the silicon oxide and silicon
nitride films underlying the polycrystalline silicon film in the range of
from 1.5 eV to 2.0 eV as shown in FIGS. 10 and 11 although most of
commercially available typical systems can measure in the range of from
1.5 eV to 2.0 eV.
It is assumed that dependencies on wavelength of reflective indexes and
damping coefficients, which have been calculated by the polarization
characteristics, are represented by functions k(.lambda.) and n(.lambda.),
respectively. These functions k(.lambda.) and n(.lambda.) are represented
by the functions n.sub.a-Si (.lambda.), n5(.lambda.), n6(.lambda.),
k.sub.a-Si (.lambda.), k5(.lambda.) and k6(.lambda.), and g, h and i
simultaneously satisfying the following formulae are determined.
k(.lambda.)=g*k5(.lambda.)+h*k6(.lambda.)+i*k.sub.a-Si (.lambda.)
n(.lambda.)=g*n5(.lambda.)+h*n6(.lambda.)+i*n.sub.a-Si (.lambda.)
g+h+i=1
FIG. 12 shows the relationship between coefficients and grain sizes in the
optical characteristics of h, i.e., the polycrystalline silicon film
having a thickness of 0.3 .mu.m. It can be seen from FIG. 12 that if the
value of h is greater than or equal to 0.3, the mean grain size is in the
range of from 0.25 .mu.m to 0.45 .mu.m.
Moreover, as shown in a correlation diagram of FIG. 13, the grain size is a
medium grain size ranging from 0.25 .mu.m to 0.45 .mu.m when h.gtoreq.0.3,
a large grain size of 0.46 .mu.m or more when h<0.3 and the
polycrystalline silicon film is thin, or a small grain size of 0.249 .mu.m
or less when h<0.3 and the polycrystalline silicon film is thick.
Furthermore, the reason why the value of h is different even if the
thickness of the polycrystalline silicon film is the same is that the
output of fluent of the excimer laser anneal varies in the range of
.+-.7%. In addition, a polycrystalline silicon having a large grain size
is suitable for a thin film transistor since the mobility of crystals
thereof is reasonably high. Thus, it is possible to accurately distinguish
non-defective from defective in a short time.
While the functions k.sub.a-Si (.lambda.) and n.sub.a-Si (.lambda.)
indicative of dependencies on wavelength of the refractive index and
damping coefficient of the amorphous silicon film have been used as the
third functions representing k(.lambda.) and n(.lambda.) of the
polycrystalline silicon film, the present invention should not be limited
thereto. According to the present invention, it is possible to
sufficiently inspect the polycrystalline silicon film using other
functions in place of the above functions. Moreover, even if fourth
functions and fifth functions are used, the effects of classification of
the mean grain sizes into three ranges of 0.24 .mu.m or less, from 0.25 to
0.45 .mu.m, and 0.46 .mu.m or more are the same. Furthermore, this method
can also calculate the margin of thickness of the silicon film produced by
the excimer laser anneal.
Another preferred embodiment of a method and system for inspecting a
polycrystalline semiconductor film, according to the present invention,
will be described below. This preferred embodiment uses .PSI. and .DELTA.
which represent a ratio of a reflectance of an s-polarized light to that
of a p-polarized light by tan(.PSI.).multidot.exp(i.DELTA.). Assuming that
reflection coefficients of the s-polarized light and the p-polarized light
are Rs and Rp, respectively, the values of .PSI. and .DELTA. are expressed
by the following formula.
Rs/Rp=tan(.PSI.).multidot.exp(i.DELTA.)
A method for inspecting, e.g., a polycrystalline silicon film having a
laminated structure of a silicon nitride film of a thickness of about 50
nm, a silicon nitride film having a thickness of about 100 nm and a
polycrystalline silicon film having a thickness of about 55 .mu.m on a
glass substrate, will be described below. Furthermore, the polycrystalline
silicon film is formed by the excimer laser anneal.
First, two standard samples of a polycrystalline silicon and an amorphous
silicon, which have different means grain sizes and surface
irregularities, are prepared. Preferably, these standard samples have the
same structure as that of the sample to be inspected. This preferred
embodiment also uses a polycrystalline silicon film, which has a laminated
structure of a silicon nitride film having a thickness of about 50 nm, a
silicon oxide film having a thickness of about 100 nm and an amorphous
silicon film having a thickness of about 55 nm and which is formed on a
glass substrate by the excimer laser anneal method using a XeCl excimer
laser.
In this preferred embodiment, laser beams having irradiation energies of
about 342 mJ/cm.sup.2 and about 305 mJ/cm.sup.2 are irradiated 26 times,
respectively, and it is assumed that the polycrystalline silicon films
have mean grain sizes of about 0.52 .mu.m and about 0.31 .mu.m, and the
surface irregularities are about 22 nm and about 53 nm, respectively. In
addition, a sample having a laminated structure of a silicon nitride film
having a thickness of about 50 nm, a silicon oxide film having a thickness
of about 100 nm and an amorphous silicon film having a thickness of about
55 nm on a glass substrate is prepared.
Then, the polarization characteristics of these standard samples are
measured by means of a spectral ellipsometer. Then, the polarization
characteristics are analyzed to calculate .PSI. and .DELTA. of the
amorphous silicon film and the respective polycrystalline silicon films.
The dependencies on wavelength of .PSI. of the obtained standard samples
are expressed by functions .PSI..sub.a-Si (.lambda.), .PSI.5(.lambda.) and
.PSI.6( .lambda.), and the dependencies on wavelength of .DELTA. thereof
are expressed by functions .DELTA..sub.a-Si (.lambda.), .DELTA.5(.PSI.)
and .DELTA.6(.PSI.).
Then, the polarization characteristics of the polycrystalline silicon
sample, the grain size of which is to be measured, i.e., the sample to be
inspected, which has a laminated structure of the silicon nitride film of
about 50 nm, the silicon oxide film of about 100 nm and the
polycrystalline silicon film of about 55 nm on the glass substrate, are
measured by means of a spectral ellipsometer. It is assumed that the
dependencies on wavelength of .PSI. and .DELTA. calculated from the
polarization characteristics are represented by functions .PSI.(.lambda.)
and .DELTA.(.lambda.), respectively. These functions .PSI.(.lambda.) and
.DELTA.(.lambda.) are represented by the functions .PSI..sub.a-Si
(.lambda.), .PSI.6( .lambda.), .DELTA..sub.a-Si (.lambda.),
.DELTA.5(.lambda.) and .DELTA.6(.lambda.), and g, h and i simultaneously
satisfying the following formulae are determined.
.PSI.(.lambda.)=g*.PSI.5(.lambda.)+h*.PSI.6(.lambda.)+i*.PSI..sub.a-Si
(.lambda.)
.DELTA.(.lambda.)=g*.DELTA.5(.lambda.)+h*.DELTA.6(.lambda.)+i*.DELTA..sub.
a-Si (.lambda.)
g+h+i=1
Also in this preferred embodiment using different functions, similar to the
correlation diagram of FIGS. 13 and 14, the grain size is a medium grain
size ranging from 0.25 .mu.m to 0.45 .mu.m when h.gtoreq.0.3, a large
grain size of 0.46 .mu.m or more when h<0.3 and the polycrystalline
silicon film is thin, or a small grain size of 0.24 .mu.m or less when
h<0.3 and the polycrystalline silicon film is thick.
The mean grain sizes of the standard samples should not be limited to those
in the above preferred embodiment, but the mean grain sizes may be 0.4
.mu.m and 0.6 .mu.m, or 0.3 .mu.m and 0.7 .mu.m. Thus, although any
combinations of mean grain sizes may be optionally selected, at lease one
standard sample preferably has a refractive index, a damping coefficient,
.PSI. or .DELTA. in the range of a mean grain size of from 0.2 .mu.m to
0.5 .mu.m. Because the surface morphology in the case of a mean grain size
of from 0.2 .mu.m to 0.5 .mu.m is worse than those in the case of other
mean grain size ranges to increase the sur | | |
