semi合集-English.pdf - 第7511页

SEMI MF1724-1104 © SEMI 2004 6 analysis steps are carried out i n a clean room and laminar flow hood, with analysts in full clean room attire. A minimum of a Class 6 clean room, as established in ISO 14644– 1 (Note 1), i…

SEMI MF1724-1104 © SEMI 2004 5

Table 1 Sample Cup Loading Protocol with Function of Calibration and Control Standards

Sample Cup Function

Acid Blank

5-ppb calibration standard

10-ppb calibration standard

20-ppb calibration standard

To provide a 4-point calibration curve in the range of the analyte and set the 5-ppb

instrument absorption value.

Two 10-ppb digested control standards To monitor the dissolved metals retention of the method and quantify variations in the

method.

Two digested blanks, acid etch mixture To monitor reagent purity and purity of the environment and to provide a zero reference

for the lot sample.

Lot samples To determine concentrations of the analytes.

10.2.5 Prepare two acid blanks and digest along with

the lot samples. Measure these blanks to provide a

quantification of variations in reagent purity and

contamination that occurs from the room environment

during processing. Enter the values for the blanks into

a database and establish statistical rules to determine

any excessive variation in the current value. Repeat the

analysis if the current value is out of statistical control.

10.3 For purposes of interlaboratory correlation and

comparison of techniques, definitions of instrument

variation and this test method variation and detection

limit are as follows:

10.3.1 Measure the instrument variation by analyzing a

2-ppb standard for 15 runs, calculating the standard

deviation of the 15 values, then calculating the 3-sigma

value for each element. Multiply the 3-sigma value by

the dilution factor to define the instrument variation.

NOTE 4: Concentration of standards are chosen based on

instrument manufacturer recommendations for signal/noise

ratio and sensitivity values.

10.3.2 Measure the method variation by analyzing a

minimum of 30 values for the 10-ppb digested control

standards. Calculate the standard deviation for these 30

values and multiply the 3-sigma value by the dilution

factor for each element to define the method variability

of this test method.

NOTE 5: Method variation can be reduced by using lower

concentration control standards and by making improvements

in the procedure, room cleanliness, and acid purity. These

improvements are monitored by continually tracking the 3-

sigma values for the digested control standards.

10.3.3 Measure the detection limit by analyzing a

minimum of 30 values for the digested acid blanks.

Calculate the standard deviation for these 30 values and

multiply the 3-sigma value by the dilution factor for

each element to define the detection limit of this test

method.

11 Procedure

11.1 Clean bottles, lids and tongs between analyses.

Conduct additional cleaning and analyses of blanks

when using new bottles, when blank values indicate

contamination, and when performing referee analyses.

Prepare clean bottles and lids as follows:

11.1.1 Rinse three times with DI water.

11.1.2 Fill the bottles with 500 mL of the acid cleaning

mixture and heat on the hot plate, lids fit loosely, for 6

h at 100°C; refill the bottles with the acid cleaning

mixture and heat on the hot plate, no lids, for 6 h at

100°C.

11.1.3 Rinse bottles and lids three times with DI water;

11.1.4 Fill bottles with 250-mL acid etch mixture, no

lids, and heat at 130 to 150°C until dryness, about 10 h.

11.1.5 Dissolve residue with 2 mL of 5% HNO

and 8

mL DI water; analyze using the GFAAS instrument.

11.1.6 Repeat entire procedure if the GFAAS analysis

shows contaminants are present in the sample bottles.

11.1.7 Triple rinse bottles and lids.

11.1.8 Fill bottles with acid cleaning mixture and heat

for 6 h at 100°C.

NOTE 6: Other cleaning procedures may be used, but must

be monitored to prove effectiveness.

11.2 Choose one 5-kg bag of polysilicon from a lot to

be sampled. Deliver the bag to the clean room for

sampling. Choose six chunks from the bag as the lot

sample. At least three of the six chunks must have an

outside surface. Each chunk shall be about 50-g

weight, with an approximate size of 3 by 3 by 3 cm,

total weight about 300 g. If the sample must be taken

in a location other than the analytical laboratory, seal

the sample in double bags and send to the laboratory.

To minimize contamination from the room environment

and the analyst, all sample collection and

SEMI MF1724-1104 © SEMI 2004 6

analysis steps are carried out in a clean room and

laminar flow hood, with analysts in full clean room

attire. A minimum of a Class 6 clean room, as

established in ISO 14644–1 (Note 1), is required to

minimize interferences from environmental sources. In

the laboratory, open the double bag according to

standard clean room practice and transfer the chunks to

a clean, numbered PTFE bottle and weigh to two

decimal places. Add about 250 mL of acid etching

mixture to each bottle to cover the chunks and seal with

PTFE lids.

11.3 Place the sealed bottles onto the hot plate in the

fume hood and heat for about 60 min at about 70°C.

Remove from heat and cool, then remove each chunk

with PTFE tongs, rinsing the surface with DI water into

the bottle. Return the etchant in the open bottle to the

hot plate and heat to dryness at 110 to 150°C.

NOTE 7: A microwave oven may be used in place of a hot

plate to reduce the digestion time required to take to dryness.

11.4 Remove the bottle from the hot plate, replace the

lid and leave to cool. Add 2 mL of 5% HNO

to the

dried etchant residue and let stand for about 20 min to

dissolve all salts. Add 8 mL of DI water, replace lid,

and swirl to mix. No solids should be observed in this

solution.

11.5 Prepare 5-ppb, 10-ppb, and 20-ppb calibration

standards, control standards, and blanks (see 10.2).

11.6 Load samples and standards into the GFAAS

instrument sample carrousel tray. Each run shall

include the following:

 Acid blank,

 5-ppb calibration standard,

 10-ppb calibration standard,

 20-ppb calibration standard,

 First 10-ppb digested control standard,

 Second 10-ppb digested control standard,

 First digested blank,

 Second digested blank, and

 Lot samples.

11.7 Choose the analysis pattern so that standards and

blanks are run before and after samples in order to

detect any shift in calibration during the analysis.

11.8 Analytical Conditions

11.8.1 Choose and record the furnace and instrument

parameters, according to the instrument manufacturer's

instructions, for the sample measurement. This

includes the following:

 Furnace time/temperature profile,

 Slit width,

 Sample injection size, and

 Wavelength.

12 Calculation

12.1 Calculate the results as follows:

DFBIM 





)( (1)

where:

M = concentration of analyte, ppbw,

I = instrument reading of analyte, ppbw,

B = instrument reading of average of two blanks,

ppbw, and

= dilution factor, final volume of acid extract

divided by the polysilicon sample weight = 10

mL  sample weight in grams.

13 Report

13.1 Report the following information:

13.1.1 Polysilicon lot sample identification,

13.1.2 Date,

13.1.3 Manufacturer, type, and model of instrument,

13.1.4 Location of laboratory and analyst,

13.1.5 Analyte values in ppbw,

13.1.6 Blank values in ppbw,

13.1.7 Weight of polysilicon sample, and

13.1.8 Confirmation of calibration standards in control.

14 Precision and Bias

14.1 Precision — The precision was determined by

making several analyses of a calibration standard to

measure within-laboratory variation and by conducting

interlaboratory correlations where polysilicon lots were

analyzed.

14.2 Within-laboratory Precision — Study of within-

laboratory variation consisted of daily analyses of

standards and lot samples over a one-year period using

the acid mixture, procedure, conditions, and statistical

control methods described in this test method.

14.2.1 Variation of the GFAA instrument used was

determined by analyzing a 2-ppb standard of each

analyte for 15 times. The instrument reading is

multiplied by the dilution factor, 10/300. This set of

analyses was repeated over a period of several months.

The standard deviation of this set of analyses represents

the variation due to instrument drift, analyst technique,

and room environment.

SEMI MF1724-1104 © SEMI 2004 7

14.2.2 Method variation was determined by analyzing

the 10-ppb digested control standards over a period of

one year. Instrument readings of concentration values

are about 10 g/L. The standard deviation of this set of

values represents the variation due to metals retention

efficiency, reagent purity, apparatus purity, environ-

ment purity, and analyst technique.

14.2.3 Lot sample analysis variation was determined

by analyzing polysilicon lot samples over a one-year

period. The standard deviation of this set of analyses

represents the variation due to sampling techniques,

contamination during sampling and processing of lots,

and the total analytical method variation.

14.2.4 The standard deviations in these studies are

summarized in Table 2.

Table 2 Standard Deviation (ppbw) Values for a

One-Laboratory Study of Variations for a One-Year

Period

Analyte

Instrument

Variation

Method

Variation

Lot Sample

Variation

Sodium 0.01 0.08 0.15

Aluminum <0.01 0.10 0.28

Iron <0.01 0.10 0.13

Chromium <0.01 0.05 <0.01

Nickel <0.01 0.03 <0.01

Zinc <0.01 0.08 0.13

14.3 Interlaboratory Variation — Interlaboratory

correlation studies were conducted to test the ability of

the method to analyze surface contamination in

polysilicon lots at the sub-ppbw level. Each laboratory

used different polysilicon sample weights, acid

mixtures, and dilution factors, but all performed the

analyses in clean rooms with high purity reagents, and

calibration controls. All polysilicon samples were

supplied by Laboratory A, with lots chosen to be

representative of polysilicon with only trace

contaminants, at sub-ppbw levels. No samples with

added contaminants were supplied; this correlation

study was to determine variation at these levels due to

sampling, the different analytical procedures, and

different laboratory environments. Laboratory A

followed this test method, while Laboratories B, C, D,

and E used variations of this test method and the acid

mixture and dilution factor for these laboratories were

not reported. The ability of the method to detect values

above the sub-ppbw level is determined by the recovery

of the control standards as discussed in Section 14.2.

Requested weights of samples were taken from a

polysilicon lot, sealed in double polyethylene bags, and

sent to the labs for analysis.

14.3.1 Laboratory A and Laboratory B used 300-g

sample weights. Five polysilicon lots were analyzed.

Detection limits for Laboratory A are based on 3 sigma

of the method variation listed in Table 2.

14.3.2 Laboratory C used 200-g sample weights. Five

polysilicon lots were analyzed.

14.3.3 Laboratory D used 200-g sample weights.

Seven polysilicon lots were analyzed.

14.3.4 Laboratory E used 80-g sample weights. Three

polysilicon lots were analyzed.

14.3.5 In each case, the average value and standard

deviation are given for each analyte reported. Values

less than the detection limit are reported as <“detection

limit” and no standard deviation is reported. Results

are tabulated in Table 3.

14.4 Bias — Although liquid standard samples are

available, no reference materials of contaminated

silicon are available to calibrate this measurement;

therefore, no bias statement is possible.

15 Keywords

15.1 acid extraction; contamination; graphite furnace

atomic absorption spectroscopy; metals; polycrystalline

silicon; surface contamination

Table 3 Surface Metals Analysis Comparison (ppbw) for a Five-Laboratory Correlation Study

Laboratory Sodium Aluminum Iron Chromium Nickel Zinc

Average <0.24 <0.30 <0.30 <0.15 <0.09 <0.24 Laboratory A

Standard Deviation … … … … … …

Average 0.23 0.27 0.03 <0.01 <0.06 0.09 Laboratory B

Standard Deviation

0.03 0.04 0.02 … … 0.03

Average 0.04 0.02 0.11 0.02 0.02 0.04 Laboratory C

Standard Deviation

0.01 <0.01 0.07 <0.01 0.01 0.01

Average 0.18 … 0.12 0.02 0.02 0.20 Laboratory D

Standard Deviation

0.07 … 0.07 0.02 0.02 0.05

Average <0.20 <0.25 <0.10 <0.13 <0.20 0.12 Laboratory E

Standard Deviation … … … … …

0.06