semi合集-English.pdf - 第778页

SEMI E104-0303 © SEMI 2000, 2003 14 APPENDIX 2 CALIBRATION NOTES NOTICE : The material in this appendix is an official part of SEMI E104 and was approved by full letter ballot procedures on July 28, 200 0 by the European…

SEMI E104-0303 © SEMI 2000, 2003 13

APPENDIX 1

PERFORMANCE TESTING

NOTICE: The material in this appendix is an official part of SEMI E104 and was approved by full letter ballot

procedures on July 28, 2000 by the European Regional Standards Committee.

A1-1 The definitions of parameters and measurement

of the terms described in SEMI E10 applies also to

LPPDs and combined LPPD/process equipment.

SEMI E104-0303 © SEMI 2000, 2003 14

APPENDIX 2

CALIBRATION NOTES

NOTICE: The material in this appendix is an official part of SEMI E104 and was approved by full letter ballot

procedures on July 28, 2000 by the European Regional Standards Committee.

A2-1 Size Calibration

A2-1.1 Size calibration is performed with

monodisperse particles with known sizes and known

optical properties. The procedure establishes the

voltage response of the LPPD for these particles. The

reported size for unknown particles is the same size of

the monodisperse particle whose voltage response is the

voltage pulse produced by the unknown particle (called

Optical Equivalent Size).

A2-1.2 If an LPPD is able to detect all of the

distribution (see counting efficiency) generated by a

monodisperse aerosol and the size of the particle

standard is known, the calibrating of an LPPD without a

reference particle counter is possible. Ideally, the data

collected by the LPPD observing a monodisperse

aerosol would describe a Gaussian particle size

distribution. However, the reported voltage pulse height

distribution is often not symmetric. Therefore, the mode

and modal values of the pulse height distribution are

not equal. The thresholds for different particle sizes are

set as the mode or median value of the observed

distribution. The modal voltage method is commonly

used.

A2-1.3 The intensity of the scattered light detected by

a photodetector depends on intensity, polarization state,

and wavelength of the incident light beam, diameter,

shape, and refractive index of the particle, as well as on

the geometrical layout of the collection optics and

detector. In the particle size range near the wavelength

(0.1λ < d

< 10λ), large oscillations can be seen in the

intensity curve of the light scattered by spherical

particles as a function of all these parameters.

Therefore, the response curve might not be monotonic.

Particles of more than one size will produce the same

voltage output signal. The calibration particles should

be chosen so that their response is not included in such

reversals of the calibration curves.

A2-1.4 For calibration checks, a reference particle

counter (RPC) with a good sizing capability could be

used. The counting efficiency of the RPC shall be 100%

for the particle sizes used in test. The counts of the RPC

and the LPPD under test in the corresponding channel

should be normalized to a standard detection area

(mm

A2-2 Particle Size Resolution

A2-2.1 Particle size resolution of a particle detector

describes its capability to differentiate between particles

of nearly the same size, or it is a measure of the range

of sizes which the counter would assign to a particular

particle if its size was determined repeatedly. The

resolution is specified by the coefficient of variation (in

%) obtained by dividing the portion of the standard

deviation of the distribution σ

LPPD

that is contributed by

the LPPD by the mean particle size.

A2-2.2 The employed particle size should be at least 2

times larger than the lower detection limit of the

instrument and is within a monotonic response range of

the LPPD response curve. The size d

and the standard

deviation σ

d,p

of the used particles is necessary to

calculate the resolution.

A2-2.3 To determine the resolution, the standard

deviation of the observed particle distribution σ

p,LPPD

calculated. The quality of the photodetector is checked

by determining the standard deviation of the pure

sensor signal σ

V,LPPD

at the analog calibration output

with a pulse height analyzer. The σ

LPPD

is calculated by

the following formula:

(

)()

, pdLPPDpLPPD

σσσ

−=

A2-3 Zero Counting

A2-3.1 The intent of this procedure is not to adjust the

thresholds for the zero count rate but to verify whether

the LPPD is within its specification. It is assumed that

the zero count level of a correct operating LPPD is

sufficiently better than its specification. Failure of the

verification test is due to a physical failure within the

LPPD and not to statistical variation in the

measurement.

A2-3.2 The output of a photodetector and the

electronic circuits of the particle detector is afflicted

with noise. The zero count rate verification is carried

out to ensure that data, especially near the detection

limit of the LPPD, is generated by particles rather than

by noise. To achieve this goal, the basic thresholds for

the LDL are to be set so that a signal-to-noise ratio of at

least 3 dB is ensured. The sampling time shall be long

enough to provide adequate sampling statistics.

SEMI E104-0303 © SEMI 2000, 2003 15

A2-4 Counting Efficiency

A2-4.1 The counting efficiency for a specific particle

size is defined as the ratio of the detected concentration

of particles to the concentration actually present. The

determination of the counting efficiency of an LPPD

requires particles of known size and concentration in

the aerosol. The counting efficiency is affected by

several factors.

• First, the counting efficiency is dependent on a

specific particle size.

• The effect of particle concentration is addressed

under zero counting and coincidence.

• Due to inhomogeneous light intensity within the

detection area of the particle counter, not all small

particles near the lower detection limit of the

instrument are detected.

• If the sampling flow is not completely contained

within the defined detection area, then some

portions of the aerosol will not be counted.

A2-4.2 Ideally, the counting efficiency of a particle

counter covering 100% of the line cross-section area as

detection area would consist of a step function at the

point of the LDL. Real particle detectors have a gradual

transition (efficiency curve) instead of a step function.

The point with 50% detection probability of all particles

of a given size moving through the detection area is

often used as a reference point. The corresponding

particle size is called the minimum detectable particle

size (see sensitivity). If the slope of the efficiency curve

for two detectors with the same specified 50%

efficiency point is different, the two detectors may

show different particle counts for the same polydisperse

aerosol. To define an efficiency curve, more than one

particle size, e.g. 5 particle sizes around the LDL

covering 0% to 100% efficiency, should be used.

Acceptable counting efficiency for single particle

counting instruments is 50% ± 20% at the minimum

detection size and 100% ± 10% for all particles larger

than 1.5 times the minimum size.

A2-4.3 The detection area of most LPPDs comprises

only a small portions of the cross-section area of the

line. Therefore, it is not possible to specify a 50%

efficiency point. For this, the commonly used definition

for sensitivity could not be used for such instruments.

A2-4.4 The actual particle concentration is determined

with a reference particle counter with a 100% counting

efficiency for the employed particle size. It is important

that there are no differences between the concentration

within the detection area of the LPPD and the

concentration monitored by the RPC.

A2-5 Sensitivity

A2-5.1 The particle size corresponding to the point

with 50% counting efficiency of the particles of a given

size is defined as the minimum detectable particle size.

This is valid only for particle counters capable to reach

100% counting efficiency for a specific particle size.

This does not apply to most LPPDs.

A2-6 Particle Concentration Effects

A2-6.1 All optical particle detectors are able to operate

accurately only within a limited range of particle

concentration. It is not the scope of this document to

define how this limit is measured or how to verify the

specification.

A2-6.2 High Concentration Effects

A2-6.2.1 The upper limitation of the particle

concentration is chiefly based on coincidence effects.

A2-6.2.2 One effect is optical coincidence. If the

particle concentration is too high, more than one

particle is present within the detection area of the

instrument. The reported particle concentration will be

less than the true value, and the reported particle size

distribution will be shifted towards the indication of

larger particles than in reality.

A2-6.2.3 The other effect is electronic coincidence. It

is defined as the inability of the electronic pulse

processing system to detect and size individual pulses

that are too closely spaced. If there are so many pulses

that they cannot completely return to the baseline, they

become superimposed. Electronic saturation occurs.

Electronic coincidence introduces errors in both particle

size and counts as well. Normally, this problem is not

critical for modern electronic systems. Optical

coincidence can become a problem at particle levels

well below the point where electronic saturation occurs.

A2-6.2.4 Another error can occur when high

concentrations of particles just smaller than the lower

detection limit of the instrument are present. Even

though no single one of these particles will be detected,

scattering-light levels from these particles can increase

the background optical noise level. So, errors might be

produced in particle count data in the lower particle size

ranges reported by the counter.

A2-6.3 Low Concentration Effects

A2-6.3.1 It is obvious that it is necessary to collect

sufficient data to determine number and size of particles

within an acceptable confidence limit.