semi合集-English.pdf - 第5363页
SEMI M54-0304 © SEMI 2003, 2004 4 5.4 The conce ntration of C incorporated on the As lattice site, [C AS ], is measured using the lo cal vibrational mode (L VM) absorption with Fourie r Transform Infrared Spectro scopy (…
SEMI M54-0304 © SEMI 2003, 2004 3
strongly depend on the post growth annealing
procedures (see Section 4.4).
4.10 To obtain SI GaAs, complete ionization of
shallow donors and acceptors and partial single
ionization of EL2 must be achieved, requiring that
[EL2] > [A] - [D] > 0 (1)
The Fermi level is then pinned approximately at ∆E
below the conduction band.
4.11 The resistivity of SI GaAs is given by
ρ = (e [n] µ
n
+ e [p] µ
p
)
-1
(2a)
where e is the electron charge, [n] and [p] the
concentration of electrons and holes and µ
n
, µ
p
the
respective mobilities. For material with ρ < 5 × 10
8
Ωcm and meeting condition (1), the second term may
be neglected, hence
ρ = (e [n] µ
n
)
-1
(2b)
The electron concentration is given by
[n] ∝ [EL2
0
] / [EL2
+
] exp ( -∆E / kT) (3)
where T is temperature, k is Boltzmann' s constant and
the concentration [EL2
+
] of the singly ionized defect is
given by
[EL2
+
] = [A] - [D] (4)
Equation (3) may be used to normalize resistivity data
ρ
M
taken at a measurement temperature T
M
. The
normalized resistivity ρ
S
at a standard reference
temperature T
S
is calculated according to
ρ
S
= ρ
M
exp (α ∆T) (5)
where ∆T = T
M
- T
S
(6)
and α depends on ∆E at T = 0 K (0.75 eV) and on T
S
.
For T
S
= 296 K (23° C), one has α = 0.0994.
4.12 Due to ionized impurity scattering µ decreases
with the concentration [I] of ionized centers, given by
[I] = 2 [A] = 2 ( [EL2
+
] + [D]
). (7)
Relations (3) and (7) imply that µ decreases with
increasing resistivity. Relation (7) further implies that,
for a given resistivity, low [D] is desirable to maximize
µ .
4.13 SI GaAs ingots usually contain dislocations
generated by thermal stress during crystal growth and
postgrowth annealing. The dislocation density (DD),
although a volume property, is evaluated by measuring
the area density of dislocations threading a substrate
surface. Structural etching with molten KOH generates
characteristically shaped etch pits at the threading
points. Hence the DD is characterized by quoting the
etch pit density (EPD).
4.14 Depending on the growth method and the ingot
diameter, the EPD may vary from essentially zero up to
2 × 10
5
cm
-2
. The lateral variation of the EPD generally
forms a pattern of globular dislocation-free “cells”
surrounded by high DD “walls”. The cell dimensions
are on the order of 100 µ m ∅ for LEC grown material
and about an order of magnitude larger for VGF grown
material.
4.15 State-of-the-art SI GaAs substrates do not contain
polycrystalline structure or twins.
4.16 The intensity of band-to-band and shallow
donor/acceptor-related radiative carrier recombination
luminescence depends on the minority carrier lifetime,
which in turn is related to the distribution of NRRCs.
These centers are generally believed to be intrinsic
defects and appear to have a minor influence on the
electrical properties.
4.17 The quality of the front surface of the wafer is
mainly determined by global and local flatness (not
addressed here in detail), surface contamination, light
point defect (LPD)
3
density and micro-roughness
(haze). Other surface irregularities to be considered in
a supplier-purchaser agreement include stain, scratches,
pits, orange peel and dimples as defined in SEMI M10.
4.18 The thickness and structure of the oxide on the
surface of SI GaAs wafers to be used for epitaxy may
be prepared to allow layer deposition without chemical
pre-cleaning by the user. Generally this material
property is guaranteed for a certain time interval only.
5 Parameter Verification
5.1 The material specification of SI GaAs according to
Section 6 requires agreements on specification
verification. Existing standard test methods are
suggested and referenced in Section 3.
5.2 If a standard test method is not available,
alternative verification methods are recommended and
described in order to identify practicable
characterization procedures that are consonant with
industry practice and cost considerations.
5.3 Some characterization methods, while intensively
used for exploratory material investigations, yield
qualitative information only and are, therefore,
inadequate for a specification verification. They are
nevertheless included, with appropriate comment, to
ensure comprehensiveness of this guideline and to
facilitate respective supplier-purchaser agreements.
3 For silicon wafers, and sometimes for GaAs, the acronym LLS
(localized light scatterer) is used.
SEMI M54-0304 © SEMI 2003, 2004 4
5.4 The concentration of C incorporated on the As
lattice site, [C
AS
], is measured using the local
vibrational mode (LVM) absorption with Fourier
Transform Infrared Spectroscopy (FTIR). Presently
two standard test methods (SEMI M30 and DIN 50449-
1) are available. SEMI M30 is valid for measurement
at room temperature and [C] > 10
15
cm
-3
. It requires
individual instrument calibration with a set of
secondary reference samples. DIN 50449-1 defines
standardized FTIR measurement parameters and
calibration factors for measurement at room
temperature and at 77K, the latter having a detection
limit [C] ≥ 10
13
cm
3
. For practically relevant
concentrations (see Section 4.7) there is no evidence for
carbon incorporation other than on the As lattice site.
Hence the calibration factors quoted in SEMI M30 and
DIN 50449-1 imply that [C] = [C
As
]. The results
obtained when using the two standard test methods
presently are conflicting, hence a harmonizing activity
is considered necessary.
5.5 The electrical resistivity ρ can be evaluated with
contacting and noncontacting techniques. The
contacting van der Pauw measurement is performed
according to ASTM F76 or SEMI M39, the latter
specifically addressing the measurement of SI GaAs.
The noncontacting measurement using a capacitive
probe is performed according to DIN 50448. This
technique enables measurement of lateral variations of
ρ. Resistivity will be quoted for an agreed standard
temperature; 296K (23° C) is recommended.
Measurements done at a different temperature must be
normalized to the standard temperature as described in
Section 4.11. The difference between the standard and
measurement temperatures must not exceed 5° C.
5.6 The electron mobility µ is measured using the Hall
effect and a van der Pauw structure. The evaluation is
performed according to ASTM F76 or SEMI M39.
5.7 The etch pit density (EPD) of LEC grown SI GaAs
is evaluated according to ASTM F1404 or DIN 50454-
1. The documents describe the etching procedure and
define test location plans linked to the crystallographic
axes.
5.8 The EPD of VGF, VCZ and VB grown SI GaAs is
10
4
cm
-2
or below. The evaluation is performed
according to DIN 50454-1 or SEMI M36. The test
locations are defined by a fixed grid (SEMI M36) or by
an adaptable grid generated by a standardized
procedure (DIN 50454-1). Guidelines are given to
assess the lateral variation of EPD.
5.9 The light point defect (LPD) density is a measure
for the density of surface irregularities that are above a
certain size limit (usually 300 nm in diameter). These
irregularities may be foreign material deposited on the
wafer surface (particles) or so-called crystal originating
pits (COPs) caused by volume material
inhomogeneities generated during the growth and/or
annealing procedures and can be revealed by
application of a light scattering technique. Depending
on the mechano-chemical polishing processes and the
light scattering technique employed, COPs may or may
not be observed. At present no standard test method for
LPD evaluation is available. Therefore, a supplier-
purchaser agreement on e.g. the used measurement
system, size restriction procedure, edge exclusion,
inclusion or exclusion of COPs is necessary. To
exclude an influence of different measuring systems the
size of LPD’s should be given in units of diameter, not
of a scattering cross section.
5.10 The concentration of the neutral double donor
[EL2
0
] is measured using the optical absorption at about
1 µ m. An absolute determination is impossible because
neither a generally accepted calibration standard nor a
standardized test procedure is presently available
4
.
Hence a supplier-purchaser agreement is necessary to
ensure reproducibility. The relative lateral variation of
[EL2
0
] can be quantitatively assessed with high
precision. The evaluation of [EL2
+
] using optical
absorption is presently considered unreliable.
5.11 The total concentration of impurities acting as
donors is measured by Spark Source or Glow Discharge
Mass Spectroscopy (SSMS, GDMS). The dominant
contributions are Si, S, O, and Te. The analytical
procedure is time-consuming and costly, hence it is
generally confined to e.g. biannual control
measurements of the supplier to ensure that raw
material supply, synthesis and crystal growth
procedures are stable. By virtue of the compensation
process (see Section 4.11) the donor concentration [D],
including intrinsic defects, is implicitly controlled by
specifying [C], [EL2] and ρ.
5.12 The total concentration of impurities acting as
acceptors is measured by SSMS or GDMS. The
analytical procedure is time-consuming and costly,
hence is generally confined to regular control
measurements of the supplier to ensure that raw
material supply and the synthesis and crystal growth
procedures are stable. The dominant contribution
usually is C, intentionally doped to control ρ. Hence it
is generally sufficient to verify that the total
concentration of acceptor impurities other than C is
small compared to [C].
5.13 The concentration of B is measured using the
local vibrational mode (LVM) absorption according to
DIN 50449-2. Established, but non-standardized
4 A DIN standard test method to measure [EL2] is in preparation and
is scheduled for publication in 2002.
SEMI M54-0304 © SEMI 2003, 2004 5
methods include GDMS, SSMS, Secondary Ion Mass
Spectroscopy (SIMS) and atomic absorption
spectroscopy (AAS).
5.14 Surface contaminants are identified and quantified
with Time-of-Flight-SIMS (TOF-SIMS) and Total
Reflection X-Ray Fluorescence (TXRF). The latter
technique is difficult to apply for impurities with atomic
weight below those of the matrix elements Ga and As.
5.15 Deposition of high quality epitaxial layers without
chemical pre-cleaning depends on the surface oxide
structure of the substrate, as prepared by the supplier, as
well as the thermal oxide desorption and the epitaxial
deposition procedures of the purchaser. Hence an
individual supplier-purchaser specific evaluation is
necessary. Test epitaxy followed by layer quality
assessment is recommended.
5.16 The lateral substrate homogeneity is assessed by
automated analytic instrumentation elaborated to
generate topographic images of important material
parameters, including electrical resistivity, EL2
concentration, EPD and minority carrier lifetime.
Macroscopic (e.g. radial) variations as well as
mesoscopic fluctuations (generally correlated to the
cellular dislocation structure) may be assessed. Such
topographic analysis is provided, respectively, by
contactless capacitive resistivity mapping (see DIN
50448), EL2 absorption topography (see footnote 3),
specular light reflection topography and
photoluminescence topography. The first three
topography techniques provide quantitative data.
Photoluminescence topography gives qualitative
homogeneity information only, because the
interrelations between the luminescence intensity and
the various radiative and nonradiative recombination
processes are only partially understood. The details of
the topographic measurements, i.e. lateral resolution
and edge exclusion, must be defined by individual
supplier-purchaser agreement.
5.17 As precipitates, other inclusions and voids are
visualized with light scattering tomography (LST). At
present it is not possible to differentiate between these
scattering centers and to obtain quantitative information
concerning their concentration and size. Hence LST
images provide a qualitative information only.
5.18 Micro-roughness, also referred to as haze, is
indicated by a diffuse reflection of collimated, laterally
extended high intensity illumination from the wafer
surface. Localized haze and haze observed on
particular wafers of a batch indicate potentially
disadvantageous variations of surface quality. The
agreement should specify the conditions of observation
(e.g. illumination intensity).
5.19 Surface irregularities (stain, scratches, pits, orange
peel, dimples) are identified by high intensity
illumination or by standard and phase sensitive
(Nomarski) microscopy. Individual agreement is
recommended concerning the permitted surface density
of these defects.
6 Ordering Information
6.1 The material properties of SI GaAs substrates, as
described in Section 4, are addressed in the ordering
agreement using the tables given below. The
subdivisions into “important”, “optional” and “other”
specifications are suggested only, i.e. supplier and
purchaser may agree on an individual choice, taking
into account the respective application as well as cost
considerations.
6.2 In addition to specifying individual material
properties as listed below, the purchase order may
specify the fabrication procedures according to Sections
4.2 and 4.3.
6.3 Substrates are usually delivered in batches. The
batch homogeneity is defined in the purchasing
agreement by variation ranges of specified parameters.
The ordering agreement shall also state whether batches
must originate from one ingot or may be assembled
from several specified or unspecified ingots. In either
case the supplier guarantees compliance with the
specified batch homogeneity by control of fabrication
technology and appropriate test sample evaluations.
6.4 Table 1 lists the parameters that are considered
important to specify the material quality of SI GaAs
substrates. The level of importance may be adjusted
individually by taking into account the intended use for
implantation or epitaxy. The quoted absolute values
suggest a customary material specification, to be
adopted unless individual considerations warrant other
choices.
6.5 Table 2 lists optional parameters that are important
for specific applications only, hence may or may not be
considered relevant for a particular supplier/purchaser
agreement. For most of these parameters, standard test
methods at present do not exist.
6.6 Table 3 lists other parameters which in general
need not be specified, but are addressed for clarification
and to provide a basis for agreement in the event that a
purchaser desires to include such parameters into a
purchasing specification.
7 Related Documents
7.1 SEMI Standards
SEMI M1 — Specification for Polished
Monocrystalline Silicon Wafers