semi合集-English.pdf - 第5430页
SEMI M59-0305 © SEMI 2005 11 5.89.1 Discussi on —These components are considered within defi ned limits of spatial wavelength (or frequency). 5.90 scan direct ion — the directi on of successi ve subsites i n a scanner si…
SEMI M59-0305 © SEMI 2005 10
5.87.4 R
qA
can also be estimated by integrating the two-dimensional power spectral density (PSD) function, PSD(f
x
,
f
y
), over the spatial frequency range between spatial frequencies that lie within the bandwidth of the measurement:
2/1
2
1
2
1
dd),(
x
x
y
y
f
f
f
f
yxyxAqA
ffffPSDR (3)
5.87.5 If the surface is assumed to be isotropic and the instrument response function is neglected, the rms
microroughness over the spatial frequency range between f
1
and f
2
can also be obtained by integrating the isotropic
PSD function:
2/1
2
1
d)(
f
f
isoqA
ffPSDR (4)
where:
2/122
2
0
)(
and),,(2
d),()(
yx
yxA
yxAiso
fff
ffPSDf
fffPSDfPSD
5.88 rms microroughness, R
q
— the root mean square of the surface profile height deviations Z(x) from the mean
line taken within the evaluation length L.
5.88.1 Discussion — R
q
is one of several statistical metrics that can be used to describe a surface profile; definitions
for other metrics and for such concepts as mean line, evaluation length, and power spectral density function, may be
found in ANSI/ASME B46.1 and ISO 4287/1.
5.88.2 The function R
q
is related to a one-dimensional measurement of the surface profile as follows:
2/1
0
2
d)(
1
L
q
xxZ
L
R (5)
5.88.3 The digital approximation of R
q
for a profile consisting of N equally spaced points is:
2/1
1
2
1
N
i
iq
Z
N
R (6)
5.88.4 Experimentally, the profile is always limited by the spatial bandwidth of the measurement. For a profile of
length L, consisting of N equally spaced points, the lower spatial frequency limit f
1
can never be less than 1/L and
the upper spatial frequency limit f
2
can never be greater than the Nyquist limit, N/2L. In practical cases, f
1
2/L; the
achievable value of f
2
depends on instrumental parameters.
5.88.5 R
q
can also be estimated by integrating the one-dimensional power spectral density (PSD) function, PSD(f),
over the spatial frequency range between two spatial frequencies, f
1
and f
2
, that lie within the bandwidth of the
measurement:
2/1
2
1
d)(
f
f
q
ffPSDR (7)
5.88.6 In all cases, R
q
must be reported together with the lower and upper limits, f
1
and f
2
, respectively, of the spatial
frequency bandwidth over which it has been determined. Alternatively, the spatial bandwidth may be expressed in
terms of the upper and lower spatial wavelengths,
f
2
) and
f
1
), respectively.
5.89 roughness — the more narrowly spaced components of surface texture.
SEMI M59-0305 © SEMI 2005 11
5.89.1 Discussion —These components are considered within defined limits of spatial wavelength (or frequency).
5.90 scan direction — the direction of successive subsites in a scanner site flatness calculation.
5.90.1 Discussion — The scanner site flatness value obtained for a site may depend on scan direction.
5.91 scanner site flatness — the maximum subsite TIR or the maximum subsite FPD, of a site.
5.91.1 Discussion — The subsite TIR is the TIR of the portion of the subsite that falls within the FQA and within
the site; the subsite FPD is the maximum FPD of the portion of the subsite that falls within the FQA and within the
site. The reference plane is calculated using all points within the subsite that fall within the FQA.
5.91.2 Precise scanner site flatness measurement requires measurement points located closely enough to reveal the
surface topography in detail. It is recommended that the scanner site flatness be measured using a data point array
with adjacent points separated by 1 mm or less.
5.92 scratch — a shallow groove or cut below the established plane of the surface of a semiconductor wafer, with a
length to width ratio greater than 5:1.
5.93 secondary flat — a flat of length shorter than the primary orientation flat, whose position with respect to the
primary orientation flat identifies the type and orientation of the wafer.
5.94 shallow etch pits — etch pits that are small and shallow in depth under high magnification, > 200. Also
known as saucer pits (see also haze).
5.95 shape — for wafer surfaces, the deviation of a specified wafer surface relative to a specified reference plane
when the wafer is in an unclamped condition, expressed as the range or total indicator reading (TIR) or as the
maximum reference plane deviation (maximum RPD) within the specified fixed quality area.
5.95.1 Discussion — This definition is analogous to the definition of flatness, which applies to the front surface
geometry when the wafer is in the clamped condition.
5.96 site — a rectangular area, on the front surface of a wafer, whose sides are parallel and perpendicular to the
primary orientation flat or to the notch bisector, and whose center falls within the FQA.
5.97 site array — a set of contiguous sites.
5.98 site flatness — the TIR or the maximum FPD of the portion of a site that falls within the FQA.
5.98.1 Discussion — Precise site flatness measurement requires measurement points located closely enough to
reveal the surface topology in detail. It is recommended that site flatness be measured using a data point array with
adjacent points separated by 2 mm or less. It is also recommended that the data set used to calculate site flatness
have data at each site corner and along each site boundary. This makes the effective site measurement area equal to
the site size.
5.99 slip — a process of plastic deformation in which one part of a crystal undergoes a shear displacement relative
to another in a fashion that preserves the crystallinity of the material.
5.99.1 Discussion — After preferential etching, slip lines are evidenced by a pattern of one or more parallel straight
lines of dislocation etch pits that do not necessarily touch each other. On {111} surfaces, groups of lines are
inclined at 60° to each other; on {100} surfaces, they are inclined at 90° to each other.
5.100 sori — the difference between the maximum positive and maximum negative deviations of the front surface
of a wafer that is not chucked from a reference plane that is a least-squares fit to the front surface.
5.101 stain — area contamination that is chemical in nature and cannot be removed except through further lapping
or polishing.
5.101.1 Discussion — Included in this category are “white” stains that are seen after chemical etching as white or
brown streaks. Not included in this category are non-removable artifacts not caused by contaminants; such artifacts
are frequently localized differences in surface texture.
5.102 subsite, of a site — a rectangular area, L
ss
× W
ss
, on the front surface of a wafer, associated with a particular
site. The center of the subsite must be within the site. Some part of the subsite must be within or on the FQA
boundary. A subsite corresponds to the instantaneous area exposed by a scanning stepper.
SEMI M59-0305 © SEMI 2005 12
5.103 surface texture — the topographic deviations of a real surface from a reference surface.
5.103.1 Discussion — Surface texture includes roughness, waviness, and lay.
5.104 swirl — helical or concentric features that are visible to the unaided eye after preferential etch, and appear to
be discontinuous under 100× magnification.
5.105 terracing — a network of contours that are associated with pyramid-like defects on epitaxially deposited
surfaces and are related to the orientation of the surface.
5.106 thermal emf — the net emf set up in a thermocouple under conditions of zero current. Also known as
Seebeck emf.
5.107 thickness, of a semiconductor wafer — the distance through the wafer between corresponding points on the
front and back surfaces.
5.108 thickness, of an epitaxial layer — the distance from the surface of a wafer to the layer-substrate interface.
5.109 tolerance — the allowable range of a specification parameter on either side of the nominal or target value.
5.110 total indicator reading, TIR — the smallest perpendicular distance between two planes, both parallel with the
reference plane, that encloses all points on the front surface of a wafer within the FQA, the site, or the subsite,
depending on which is specified.
5.111 total thickness variation, TTV — the difference between the maximum and minimum values of the thickness
of a wafer.
5.111.1 Discussion — Initially, the TTV was determined by measurement at a small number of points, generally
five or nine, but modern measurement equipment samples the wafer at relatively small intervals over its entire
extent.
5.112 twin boundary — a coherent planar interface that separates two parts of a crystal lattice that are related to
each other in orientation as mirror images.
5.113 warp, of a semiconductor wafer — the difference between the maximum and minimum distances of the
median surface of a free, unclamped wafer from a reference plane.
5.114 waviness — the more widely spaced component of surface texture.
5.114.1 Discussion — Waviness may be caused by such factors as machine or work piece deflections, vibration, and
chatter. Roughness may be considered as superimposed on a wavy surface
6 Symbols
6.1 English Alphabetical Symbols
6.1.1 r — radial dimension of wafer coordinate system with origin at the wafer center.
6.1.1.1 Discussion — Note that the radial dimension associated with the edge profile template in SEMI M1 is called
x and that this dimension is positive away from the actual edge of the wafer.
6.1.2 t — wafer thickness.
6.1.2.1 Discussion — The wafer thickness is sometimes indicated with a capital T, but this usage is to be
discouraged because T also stands for temperature.
6.1.3 x — direction of the wafer coordinate system along the diameter perpendicular to the bisector of the primary
fiducial with origin at the center and positive direction to the right when the top surface is up and the primary
fiducial is toward the operator.
NOTE 2: See also ¶6.1.1.1.
6.1.4 y — direction of the wafer coordinate system along the diameter that is the bisector of the primary fiducial
with origin at the center and positive direction to the top (away from the fiducial) when the top surface is up and the
primary fiducial is toward the operator.