semi合集-English.pdf - 第7316页

SEMI MF1188 © SEMI 2003, 2005 10 APPENDIX 1 ORIGINAL (LONG BA SELINE) METHOD NOTICE : The material in this appendix is an official part of SEMI MF1188. Approval was by full letter ballo t procedures wi th publicat ion au…

100%1 / 7923
SEMI MF1188-0305 © SEMI 2003, 2005 9
15.2 The certification measurements were made using a short-baseline procedure similar to that in this method
except that the endpoints of the baseline were taken as the average of the values between 1030 cm
1
and 1050 cm
1
and between 1060 cm
1
and 1080 cm
1
, respectively.
15.3 This uncertainty is expected to represent the best attainable precision with this method.
16 Bias
16.1 The uncertainty in the absolute determinations of oxygen content in silicon during the Grand Round Robin
resulted primarily from the uncertainty in the chemical analysis measurements and any variations in the oxygen
content of the various slices used in the test. The first of these is estimated to be about 6%, and the second is
estimated to be much smaller.
1
This uncertainty directly affects the assigned value of the IOC-88 calibration factor.
Errors in this calibration factor cause errors in the value of the derived oxygen content. Therefore, the bias of these
measurements is estimated not to exceed 6%.
17 Keywords
infrared absorption; infrared spectrophotometry; interstitial oxygen; oxygen; silicon
SEMI MF1188 © SEMI 2003, 2005 10
APPENDIX 1
ORIGINAL (LONG BASELINE) METHOD
NOTICE: The material in this appendix is an official part of SEMI MF1188. Approval was by full letter ballot
procedures with publication authorized by the NA Regional Standards Committee on December 10, 2004.
A1-1 Introduction
A1-1.1 §§1 through 8, 10, and 11 of the main test method also apply to this method.
A1-1.1.1 If a dispersive spectrophotemeter is being used, disregard ¶7.2.
A1-2 Procedure
A1-2.1 Instrumental Checks
A1-2.1.1 Establish the 100% baseline to measure the noise level.
A1-2.1.1.1 On double beam instruments, record the transmittance spectrum with both the sample and reference
beams empty.
A1-2.1.1.2 On single-beam instruments, obtain the transmittance spectrum as the ratio of two spectra taken with the
sample beam empty.
A1-2.1.1.3 Plot the 100% baseline over a wavenumber range covering 900 cm
1
to 1300 cm
1
. If the baseline is not
100% ± 0.5% over the entire range, increase the measurement time until it is. If the problem persists, have the
instrument repaired.
A1-2.1.2 Applies to dispersive (DIR) instruments only. Establish the 0% line. With the sample beam blocked,
record the instrument zero over the range from 900 cm
1
to 1300 cm
1
. If a significant non-zero signal is recorded in
that range, check the instrument for stray light reaching the detector. If the problem persists, have the instrument
repaired.
A1-2.1.3 Applies to Fourier transform (FT-IR) instruments only. Record the throughput characteristics of the
spectrophotometer by plotting a single-beam spectrum, obtained with the sample beam empty, over the wavenumber
range from 450 cm
1
to 4000 cm
1
. Use such a spectrum, recorded after the instrument has been properly aligned
according to factory specifications as a reference to evaluate the instrument's performance. Whenever the spectrum
obtained deviates significantly from the instrument's reference spectrum, realign the instrument.
A1-2.1.4 Determine mid-scale linearity of the instrument by obtaining an air reference spectrum of the silicon
reference specimen over the wavenumber range from 1600 cm
1
to 2000 cm
1
. If the value of the transmittance is
not 53.8% ± 2% over this wavenumber range, align the sample with the spectrometer in accordance with ¶8.3.
A1-2.1.5 Determine the measurement time for the spectra by obtaining the transmittance spectrum of a high
resistivity (greater than 5 ·cm), 0.04 cm to 0.085 cm thick double-side polished silicon slice containing between
12 ppma and 18 ppma oxygen (IOC-88) using a minimum of 64 scans for FT-IR instruments, or, for dispersive
instruments, a speed such that the full peak height is recorded. If the ratio of the net amplitude of the oxygen band,
T
b
- T
p
, to the standard deviation in the transmittance spectrum is not greater than 100, increase the number of scans
(FT-IR) or reduce the scan speed (DIR) until that criterion is met.
A1-2.2 Surface Oxide Removal — Immediately prior to the initial measurement in any laboratory, etch all
specimens, including the reference specimen, in hydrofluoric acid to remove any surface oxide and rinse.
A1-2.3 Specimen Thickness — Measure the thicknesses of the test and reference specimens to within ± 0.2%, at
their centers. If thickness of the reference specimen does not match the thickness of the test specimen to within ±
0.5%, obtain a reference specimen with the proper thickness.
A1-2.4 Temperature — Measure and record the temperature of the spectrophotometer chamber.
A1-2.5 Infrared Transmitter Spectrum — Obtain the spectrum with a resolution of 4 cm
1
, or better, at 1107 cm
1
,
for FT-IR instruments or, for dispersive instruments, 5 cm
1
, or better, over (at least) the range from 900 cm
1
to
1300 cm
1
. The test and reference specimens must be positioned so that the IR beam is centered on them. On
SEMI MF1188-0305 © SEMI 2003, 2005 11
double beam dispersive instruments, obtain the transmittance spectrum with the oxygen-free reference specimen in
the reference beam, and the test specimen in the sample beam. On single beam instruments, compute the
transmittance spectrum as the ratio of the emission spectrum of the test specimen to the emission spectrum of the
reference specimen.
A1-2.6 Plot the transmittance spectrum over the range from 900 cm
1
to 1300 cm
1
.
A1-2.7 Define the baseline by drawing a straight line from 900 cm
1
to 1300 cm
1
. Use the average transmittance
in the gions from 900 cm
1
to 1000 cm
1
, and 1200 cm
1
to 1300 cm
1
, to define the endpoints of the straight line.
A1-2.8 Locate the wavenumber corresponding to the minimum transmittance in the region from 1102 cm
1
to 1112
cm
1
. Record the value of that wavenumber, to five significant figures, as W
p
. Record the minimum transmittance
as T
p
, the transmittance at the absorption peak. Record the baseline transmittance, T
b
, as the value of the baseline
defined in ¶A1-2.7 at W
p
. Record both T
p
and T
b
to three significant figures.
A1-2.9 Determine and record the full width at half maximum (FWHM) of the peak.
A1-3 Calculations
A1-3.1 Calculate the peak and baseline absorption coefficients using the following equations:
p
x
p
xx
p
T
eTee
x 18.0
36.0)09.0()09.0(
ln
1
70.12270.170.1
(A1-1)
b
x
b
xx
b
T
eTee
x 18.0
36.0)09.0()09.0(
ln
1
70.12270.170.1
(A1-2)
where:
p
=
peak absorption coefficient, cm
1
,
b
=
baseline absorption coefficient, cm
1
,
x = thickness, cm,
T
p
= peak transmittance, and
T
b
= baseline transmittance.
A1-3.2 Calculate the net absorption coefficient,
O
, due to interstitial oxygen:
bpO
(A1-3)
A1-3.3 Calculate the interstitial oxygen content of the silicon slice as follows:
O
28.6 atomic ppm ion,concentratoxygen alInterstiti
(A1-4)
or
O
173
1014.3 atoms/cm density, umeoxygen vol alInterstiti (A1-5)
A1-3.4 This value of oxygen content is based on the IOC-88 calibration factor. If desired, calculate the interstitial
oxygen content in terms of any other standardized calibration factor using the appropriate conversion factor in SEMI
M44.
A1-4 Report
A1-4.1 Report the following information:
A1-4.1.1 The instrument used, the operator and the date of the measurements,
A1-4.1.2 Test method and baseline used (SEMI MF1188, Long Baseline),