semi合集-English.pdf - 第3768页

SEMI F5-1101 © SEMI 1990 , 2001 12 A P PENDIX 1 END-OF-PIPE A BATEMENT DESIGN NOTE: The m aterial in this appen dix is an off icial part of SEMI F5 and was approved by full letter ballot procedures on August 27, 2001. A1…

100%1 / 7923
SEMI F5-1101 © SEMI 1990, 200111
Table 1 Exhaust System Selection Table
Process Gases/Chemicals Reason for POU
Abatement
POU
Abatement
Present
End-of-pipe Exhaust System Type
VOC Yes GeneralSolvent/strippers
No VOC
Vapor prime HMDS No Acid or ammonia
Acid baths HCl/HF/Acetic No Acid
Aqua regia/hot nitric No Acid aerosolsAcid baths
Yes Acid
NH
3
No AmmoniaAmmonia baths
Yes Acid
No AcidCorrosives
Yes Acid
No AmmoniaAmmonia
Yes Acid
Gas bottle purge
Hydrides/pyrophorics Yes General
No General
Ammonia No Ammonia
VOC No VOC
Chemical dispensing
Corrosives No Acid
HNO
3
/NO
x
Yes AcidChemical wafer thinning of
raw wafers
No NOx/acid scrubber
EPI H
2
/Hydrides/HCl Fire and safety
Solids blockage
Yes Acid, H
2
reclaim, or general
Dry etching – metal Cl and Br chemistries Blockage/Corrosion Yes Acid
Dry etching – oxide/poly/etc. Fluoride chemistry No Acid
LPCVD nitride SiH
2
Cl
2
/NH
3
Solids blockage
Fire and safety
Yes Acid
LPCVD poly/oxide SiH
4
/fluoride
chemistry
Fire and safety
Solids blockage
Yes Acid
PECVD – poly/oxide SiH
4
/fluoride
chemistry
Fire and safety
Solids blockage
Yes Acid
SiH
4
/WF
6
/H
2
Fire and safety
Solids blockage
Yes AcidPECVD nitride and tungsten
SiH
4
/NH
3
Fire and safety
Solids blockage
Yes Acid
Implant Hydrides, BF
3
Hazardous deposits Yes Acid
Ammonia No AmmoniaRTP
Yes General
NOTE T1: POU Abatement may be used for the reduction of PFCs and by-products from the process.
SEMI F5-1101 © SEMI 1990, 2001 12
APPENDIX 1
END-OF-PIPE ABATEMENT DESIGN
NOTE: The material in this appendix is an official part of SEMI F5 and was approved by full letter ballot
procedures on August 27, 2001.
A1-1 Referenced Documents
A1-1.1 Gas-Liquid Reactions
12
A1-1.2 Design and Selection of Spray/Mist
Elimination Equipment
13
A1-1.3 Absorption, Distillation, and Cooling Towers
14
A1-1.4 Perry' s Chemical Engineers' Handbook
15
A1-1.5 Absorption and Extraction
16
A1-1.6 Mass Transfer
17
A1-1.7 Mass-Transfer in Co-current Gas-Liquid
Flow
18
A1-1.8 Mass-Transfer Operations
19
A1-2 Principles
A1-2.1 Principles of gas absorption and methods of
design of gas absorption equipment are treated
extensively in chemical engineering literature (See
Sections A1-1.1–A1-1.8). The sources cited present
rigorous methods for design.
A1-2.2 Practical design of scrubbers is at least semi-
empirical; and in the case of air pollution abatement
systems, design is often directly based on experimental
data.
A1-2.2.1 Consequently, designs should be based on
mass-transfer coefficients that have been determined
experimentally under, as nearly as possible, the
conditions of expected scrubber operation.
A1-2.2.2 It should be noted that the typical
concentrations of the existing empirical data base is
several orders of magnitude greater than the typical
concentrations of corrosives exiting a semiconductor
12 McGraw-Hill, New York (1970), Danckwerts, P.V.
13 Chemical Engineering Volume 91, No. 21, pp 82-89 (October 15,
1984), Holmes, T. L.; and Chen, G. K.
14 Longmans, Green, London (1961), Norman, W. S.
15 6th Edition, McGraw-Hill; New York (1984), Perry, R. H.; and
Green, D.
16 2nd Edition; McGraw-Hill, New York (1952), Sherwood, T. K.;
and Pigford, R. L.
17 McGraw-Hill, New York (1975), Sherwood, T.K.; Pigford, R. L.;
and Wilke, C. R.
18 Chemical Engineering Sciences Volume 33, No. 12, pp1675-1680
(1978), Shilimkan, R. V. and Stepanek, J. B.
19 3rd Edition; McGraw-Hill, New York (1980), Treybal, R.E.
plant (before end-of-pipe scrubbers), therefore the
efficiencies in these data bases may not be
representative of low concentrations.
A1-2.2.3 Attempts at developing the empirical
database at these low concentrations suggest that actual
removal efficiencies are lower than those predicted by
most scrubber design texts.
A1-2.2.4 As a result, new scrubbers may need to have
additional packing to ensure theoretical removal
efficiencies are achieved.
A1-2.3 Rigorous custom design of a scrubber for a
particular service requires full knowledge of the
conditions that are to be encountered as well as of the
performance that is to be attained.
A1-2.3.1 Gas flow rates should normally be known
fairly accurately — at least when the plant is originally
designed.
A1-2.3.2 Concentrations of acid and caustic gases will
need to be estimated.
A1-2.3.3 Conditions to be met may be changed several
times over the probable life of the scrubbing equipment,
as changes are made in the semiconductor fabrication
processes and the exhaust system.
A1-2.3.4 Exhaust flow may also be increased, to
accommodate equipment additions, thus raising the
possibility the scrubber may become overloaded.
A1-2.4 Studies by SEMATECH indicate that the
concentration of corrosives in the exhaust entering the
end-of-pipe scrubber(s) range from <1 ppmv to as high
as 10 ppmv. Additionally, the typical scrubber with
99+ % theoretical removal efficiency was actually
achieving removal efficiencies ranging from 80%–95%.
A1-2.5 Scrubbers being applied to corrosive pollutants
are usually standardized units assembled by a number
of vendor companies. These units are designed to
handle specified ranges of gas flow, but in other
respects are not designed but simply built. If they are
packed-bed scrubbers, they are usually fitted with
standard depths of packing. The liquid flow rates can
be varied within limits; but otherwise, the performance
potentials are fixed unless there are provisions for
changing the depth of packing and perhaps the kind of
packing. Once the system is installed, such changes to
increase the efficiency will probably increase the
SEMI F5-1101 © SEMI 1990, 200113
pressure drop of the scrubber, which will require adding
to the fan capacity.
A1-2.6 To ensure, in advance, that the scrubber will
have adequate removal efficiency, performance tests
can be conducted. Measurements should be made of
the contaminant concentrations in typical plant exhaust
gas streams, and the resulting data should be used to set
conservative performance requirements for the
scrubber. Performance tests on the scrubber should be
conducted prior to exposing the scrubber to the
semiconductor plant exhaust, preferably at the vendor’s
facility. Performance tests should identify the critical
operating variables and optimize the settings to achieve
optimal removal efficiencies at the lowest cost.
A1-2.6.1 Onsite performance tests may be preferable,
when qualifying abatement systems, so that accurate
process conditions can be produced. Permit exemptions
may be obtained for the time-period of system
qualification.
A1-2.7 Design and construction of scrubbers are
simplified by two factors: the exhaust gas streams are at
essentially ambient temperature, and they carry few
insoluble solids. The low temperature allows use of
corrosion-resistant plastic construction. The absence of
significant solids avoids blockages in the scrubbers,
which favors use of efficient gas-liquid contactors such
as packed beds.
A1-2.8 Two of the most favored scrubber types are the
counter-current packed tower and the horizontal cross-
flow scrubber. In some instances, the choice between
these devices may be determined by plant layout. As
usually constructed, these two scrubber types have low
to moderate gas pressure drops and are primarily suited
to absorption of gases. They will also collect the
coarser mists and sprays but are unsuited to collection
of fine mists and fumes having particle sizes under 2 to
3 micrometers.
A1-2.9 In principle, the counter-current tower offers
the highest efficiency with the least consumption of
absorbent liquid. However, if untreated water is
recirculated through the tower, as is commonly done in
semiconductor plants, the advantages of the
countercurrent operation are lost. On the other hand, if
the absorbed gas is reacted with an appropriate reagent
in the water, essentially the same performance can be
obtained with either the countercurrent or co-current
operation (see Section A1-1.7). In co-current operation,
it is possible to use higher gas velocities to obtain
higher mass transfer rates so that the scrubber can be
made smaller. Gas pressure drops will also be
increased, but this may be desirable if it is necessary to
collect a fine aerosol as well as absorb a gas.
A1-2.10 Horizontal cross-flow packed-bed scrubbers
may have advantages where the installation imposes
restrictions on height or other similar layout problems.
However, in other respects, they do not offer
fundamental advantages over the packed tower
scrubbers.
A1-2.11 Under special circumstances, other types of
scrubbers may be appropriate for collection of aerosols.
High-energy scrubbers, of the venturi type, or scrubbers
with high-pressure drop, sub-micron filters (e.g., fiber
bed) are appropriate where fine aerosols must be
collected. Wherever possible, measures should be
taken to avoid formation of such aerosols; but when
they are unavoidably produced, the source should be
ventilated through a separate scrubbing system of
appropriate design.
A1-2.12 Packing in scrubbers provides a large
interfacial area between the liquid and the gas, and also
induces fluid turbulence to promote mass transfer.
Packing is commonly placed in scrubbers in a random
manner. Raschig and Pall rings as well as Berl saddles
were common early packing materials used. In recent
years, numerous proprietary packing designs have been
introduced, with the objective of providing increased
mass-transfer efficiency with reduced gas pressure
drop. The validity of claims for superior performance
has been established by the SEMATECH experiments
in Section 13.1.
A1-2.13 Numerous precautions are required when
installing and operating packed-bed scrubbers. It is
essential to attain an even distribution of liquid over the
packing; channeling of liquid or gas flow results in loss
of removal efficiency. Inadequate liquid distribution
can result in a reduction of the removal efficiency to as
little as one seventh of theoretical (See Referenced
Document A1-1.9). The predominant liquid
distribution system employed is spray nozzle style.
This style is inexpensive, however, it is difficult to
maintain and tends to entrain more water than weir-
style. Weir-style liquid distribution systems are more
expensive. However, they require less preventive
maintenance, provide equal liquid distribution over a
wide range of flow rates, and are less prone to liquid
carry-over. Precautions are treated at length in
previously cited literature.
A1-2.14 A critical aspect of any scrubber design is the
separation of entrainment. Carry-over of spray from a
scrubber can produce maintenance issues and mist
emissions. Failure to provide adequate entrainment
separation has been one of the most common problems
in scrubber installations. The predominant mist
separator employed is a mesh style. Mesh style
separators are effective on the smaller water droplets
associated with spray nozzles; however, maintenance