semi合集-English.pdf - 第7337页
SEMI MF1389-0704 © 2004 5 10 Instrument Calibration 10.1 Measurement of silicon standards ne ed only be performed onc e unless instrument characteri stics have drifted or hardware chan ges have been made. The regular use…
SEMI MF1389-0704 © 2004 4
Boron Features, cm
−
1
Phosphorus Features, cm
−
1
B
TO
(b
2
) = 8777.5
P
TO
(b
2
′) = 8790.4
B
B
TO
(b
3
) = 8763.2 P
TO
(b
1
) = 8778.0
C
B
TO
(b
4
) = 8752.1
P
TO
(b
3
′) = 8771.3
B
B
TO
(b
5
)= 8742.6 P
TO
(b
2
) = 8756.0
C
P
TO
(b
4
′) = 8756.2
B
P
TO
(b
5
′) = 8745.4
B
P
TO
(b
3
) = 8740.2
C
A
The two-electron transition is represented by 2e (see Ref (3)).
B
Beta series transition.
C
Alpha series transition.
5.1.5.1 Discussion — “I” indicates intrinsic silicon
emission, “TO” indicates the transverse optical phonon
associated with the transition, and “FE” refers to the
free exciton recombination responsible for the emission
(see Table 1 for line locations).
5.1.6 phonon — a quantum of lattice vibrational
energy, as a photon is a quantum of electromagnetic
energy.
5.1.6.1 Discussion — The recombination of excitons in
silicon requires a momentum conserving mechanism
owing to the indirect band gap of the crystal. Phonons
provide such a mechanism. The principal phonon types
of interest in silicon luminescence are the transverse
acoustic (TA), transverse optical (TO), and longitudinal
optical (LO) phonons. These test methods address the
use of features associated with the TO phonon as well
as those not including phonon emission in their
momentum-conserving processes. These latter features
are designated “NP” or no-phonon features.
6 Summary of Test Method
6.1 A sample of monocrystalline silicon is cooled to 4.2
K and photoexcited with greater-than-bandgap energy
light at one of two intensities listed, depending upon the
type of instrument used. The resulting luminescence is
collected and recorded. Spectral features corresponding
to intrinsic silicon and extrinsic impurity emissions are
measured and related to calibration curves to yield
dopant density.
7 Apparatus
7.1 Cryostat — To maintain sample temperature at 4.2
K. Either open-cycle liquid helium immersion or
exchange gas cryostats, or closed-cycle refrigeration
systems may be used. The bath immersion type
cryostat is recommended for higher confidence in the
temperature stability of the sample. In both the
exchange gas and closed-cycle systems, careful
attention must be paid to thermal sinking and accurate
sample temperature measurement (see Section 3.3).
7.2 Sample Holder — Which does not cause excessive
strain on the ample through spring forces or other
means, so as to avoid line splitting associated with
crystal stresses.
7.3 Laser Excitation Source — Capable of generating
electron hole pairs in the silicon crystal. An argon-ion
laser operated at 514.5 nm is used. To obtain accurate
measurements, laser light intensity must be controllable
and stable.
7.4 Infrared Spectrophotometer — Equipped with a
detector and optics suitable for use between 8750 and
9300 cm
−1
, and capable of at least 0.5 cm
−1
resolution at
9300 cm
−1
.
8 Reagents
8.1 Purity of Water — Reference to water shall be
understood to mean Type E-1 or better water as
described in ASTM Guide D 5127.
8.2 Nitric Acid (HNO
3
), 65%, in accordance with Grade
2 of SEMI C35.
8.3 Hydrofluoric Acid (HF), 48%, in accordance with
Grade 2 of SEMI C28.
8.4 Hydrogen Peroxide (H
2
O
2
), 30%, in accordance
with Grade 2 of SEMI C30.
8.5 Acid Cleaning Mixture, (1:1:1:25) HNO
3
:HF:H
2
O
2
:
H
2
O.
8.6 Hazards — Chemicals used in this procedure are
potentially harmful and must be handled with the
utmost care at all times.
9 Sample Preparation
9.1 Perform either the procedure in Section 9.1.1 or the
procedure in Section 9.1.2 to remove all work damage
and surface contamination on all samples except as-
received chemical-mechanically polished wafers, which
do not need further preparation.
9.1.1 Etch sample using a suitable etchant (for example,
the acid cleaning mixture of Section 8.5 or CP-5
hydrofluoric-nitric-acetic acid etchant).
9.1.2 Polish the surface of the sample with a suitable
chemical-mechanical polishing compound.
9.2 Luminescence efficiency has been observed to
decrease after etching, thus introduction of the sample
into the cryostat within a few hours of etching is
recommended. Chemical-mechanically polished
samples appear to be more stable.
2
SEMI MF1389-0704 © 2004 5
10 Instrument Calibration
10.1 Measurement of silicon standards need only be
performed once unless instrument characteristics have
drifted or hardware changes have been made. The
regular use of secondary standard samples, herein
referred to as audit samples, provides a mechanism for
instrument calibration verification and generates long-
term instrument performance statistics. At least four
standard samples should be used for each impurity
calibration curve adjustment.
10.1.1 Obtain samples with known impurity densities
and collect luminescence spectra at the appropriate
excitation intensity listed in Section 11.1 using both
standard and high resolution as listed in Section 11.2.2.
10.1.2 Find the peak intensities as described in Section
12 and find the ratios of appropriate spectral lines as
described in Section 13.1.
10.1.3 Adjust calibration data of Figure 1 or Figure 2,
corresponding to Test Method A or Test Method B,
respectively, by translating the curves such that they
intersect points corresponding to PL data of the known
samples. Make no change to the slope of the curves
because the slopes are a function of the exciton
recombination dynamics only and are independent of a
given instrument’s optical response.
10.1.4 These curves become the given instrument’s
calibration curves for the analysis conditions used.
10.2 Determine the short-term one-sigma precision of
the instrument.
NOTE 3: This procedure needs to be performed only
occasionally unless instrument characteristics have drifted or
hardware changes have been made.
10.2.1 Analyze a sample with impurity concentrations
similar to typical samples analyzed by the instrument in
normal operation nine times.
10.2.2 Find the peak intensity as described in Section
12 to obtain dopant densities.
10.2.3 Calculate the open-sigma standard deviation for
all impurities present.
11 Procedure
11.1 Excitation Conditions
11.1.1 Test Method A—High Excitation Intensity
Conditions — Set the output of the argon ion laser to
300 mW of 514.5 nm light. Pass the beam through an
infrared cutoff filter and a chopper with a 50% duty
cycle. Beam can be reflected with 2 to 3 mirrors with a
nominal 1-m distance between sample and laser. Pass
the unfocused beam through the cryostat windows. Set
the beam diameter to 2.5 mm.
DOPAN T DEN SI TY (c m
−
3
)
NOTE: The PL intensity ratio of P
TO
(BE)/I
TO
(FE) and
B
TO
(BE)/I
TO
(FE) are plotted against the P and B dopant
densities, respectively.
11
Figure 1
High Excitation Condition Impurity Calibration
Curves for Boron and Phosphorus (Test Method A)
log of dopa nt de nsit y (c m
−
3
)
NOTE: Slope = 1.0.
Figure 2
Low Excitation Condition Generic Impurity
Calibration Curve (Test Method B)
11.1.2 Test Method B—Low Excitation Intensity
Conditions — Set laser intensity such that the
luminescence is as close to the EHD onset point as
practical from signal-to-noise considerations. Collect
luminescence spectra of the sample used in Section
10.2 at several different laser beam intensities,
extending from well above to as far below the EHD
SEMI MF1389-0704 © 2004 6
onset point as permissible from signal-to-noise
considerations. Plot the EHD luminescence intensity at
8720 cm
−1
versus the laser intensity used on linear axes.
The EHD onset power can be determined by
extrapolating the nonzero EHD points down to zero
intensity using a straight line.
9
11.2 Spectrometer Parameters
11.2.1 Spectral Ranges Required (minimum)
TO region—8757–8889 cm
−1
(1142–1125 nm)
NP region—9242–9294 cm
−1
(1082–1076 nm)
11.2.2 Resolution
Standard resolution = 2 cm
−1
(or 0.2 nm)
High resolution = 0.5 cm
−1
(or 0.05 nm)
Use standard resolution when calculation of aluminum
or arsenic concentrations is not desired or not
necessary. Otherwise, use high resolution.
11.3 Audit Specimen Analysis — Keep one set of
specimens as secondary standards to periodically check
instrument performance. Make sure that at least one of
these samples reflects typical material properties for the
laboratory concerned. Analyze at least one of these
samples with each batch of unknown samples to
confirm reproducibility of analysis conditions from run
to run.
11.3.1 Collect luminescence spectra of audit specimens
before unknown samples, preferably in the same sample
holder batch using parameters listed in Sections 11.1
and 11.2.2. Reduce spectral data as described in Section
12.
11.3.2 If the audit specimen results fall outside the
acceptable control limits established for an instrument,
the laser power may be adjusted to compensate for drift
in the optical transfer efficiency of the sample
excitation subsystem. If the source of drift cannot be
ascribed to this simple mechanism, thoroughly
investigate and correct the problem. Instrument
recalibration may be required.
Figure 3
Baseline Drawing Example
11.4 Collect luminescence spectra of unknown samples
with spectrometer parameters as listed in Section 11.2.
Find the peak intensity as described in Section 12.
Figure 4
High-Resolution (0.5 cm
−1
) Spectrum of
NP (No-phonon) Region
12 Peak Intensity Determination
12.1 Use the baseline drawing method depicted in
Figure 3. For the I
TO
(FE) line, draw the baseline
between (a), the point at 8880 cm
−1
, and (b), the point