semi合集-English.pdf - 第4265页
SEMI F79-0703 © SEMI 2003 2 a halogenated hydrocarbon, is a radical specie of fluorine, chl orine or other haloge n. A low pressure plasma is t h e typical m ethod of creating this radi cal. In addition, Winters (78) poi…
SEMI F79-0703 © SEMI 2003 1
SEMI F79-0703
GUIDELINE FOR GAS COMPATIBILITY WITH SILICON USED IN GAS
DISTRIBUTION COMPONENTS
This guideline was technically approved by the Global Gases Committee and is the direct responsibility of
the North American Gases Committee. Current edition approved by the North American Regional Standards
Committee on April 11, 2003. Initially available at www.semi.org June 2003; to be published July 2003.
1 Purpose
1.1 The purpose of this guideline is to identify resource
information on compatibility of gases in contact with
silicon in the wetted path of a gas delivery system
operating at typical gas stick conditions.
2 Scope
2.1 The information and conclusions provided are
taken from published literature. References are cited.
No opinion is made as to the validity of the published
conclusions. The suggestions are specific to high purity
silicon, single or poly crystal, covered by native oxide.
NOTICE: This standard does not purport to address
safety issues, if any, associated with its use. It is the
responsibility of the users of this standard to establish
appropriate safety and health practices and determine
the applicability of regulatory or other limitations prior
to use.
3 Limitations (Prescribed Conditions)
3.1 The guideline applies to the gas distribution system
for a typical semiconductor process tool. It is assumed
that silicon is in the wetted path as a coating or a
silicon-based component (eg. a valve, orifice, sensor,
mass flow controller) in the gas distribution system.
The ambient temperature and operating temperature
limits are typical for the industry. No additional energy
source, thermal, radiative or ionizing, is present. The
moisture content of the gas is not specified for the
purpose of this guideline; the recommendations apply
whether or not the conditions are anhydrous.
4 Referenced Standards
4.1 SEMI Standard
SEMI E52 — Practice for Referencing Gases Used in
Digital Mass Flow Controllers
NOTICE: Unless otherwise indicated, all documents
cited shall be the latest published versions.
5 Classification of Gases
5.1 The gases included are taken from SEMI E52; the
number assigned to a gas is the same as in that
document.
5.1.1 Inerts
5.1.2 Hydrogen
5.1.3 Hydrocarbons
5.1.4 Halogenated Hydrocarbons
5.1.5 Hydrides
5.1.6 Halogens, other than Fluorine
5.1.7 Halides, other than Fluorides
5.1.8 Fluorine and Fluorides
5.1.9 Organo-metallic and Siloxanes
5.1.10 Oxygen and Oxides and Sulfides
5.1.11 Nitrogen and Nitrogen Compounds
5.1.12 Acids
5.1.13 Other
6 Review of the literature
6.1 Inerts
6.1.1 References (1, 2, 3, 4, 5) are cited which indicate
these gases do not react with silicon under the
prescribed conditions.
6.2 Hydrogen
6.2.1 References (1, 2, 3, 4, 5) are cited which indicate
these gases do not react with silicon under the
prescribed conditions. Hydrogen has been reported to
react with silicon (75) and silicon carbide (76) at
temperatures above 1100° C.
6.3 Hydrocarbons
6.3.1 References (1, 2, 3, 4, 5, 55, 56) are cited which
indicate these gases do not react with silicon under the
prescribed conditions.
6.4 Halogenated Hydrocarbons
6.4.1 References (1, 2, 3, 4, 5, 35, 51, 54, 55, 56, 57,
58, 59) are cited which indicate these gases do not react
with silicon under the prescribed conditions. Extensive
review papers by Flamm (55) and Coburn (56) cite
numerous references on the etching of silicon and other
materials using halogenated hydrocarbons. The critical
requirement for etching of silicon, or any material using
SEMI F79-0703 © SEMI 2003
2
a halogenated hydrocarbon, is a radical specie of
fluorine, chlorine or other halogen. A low pressure
plasma is the typical method of creating this radical. In
addition, Winters (78) points out that for an
“unassisted’ etching reaction to take place adsorption
onto the silicon surface must be followed by four well
defined steps. Should any of the steps not occur, or be
interrupted, etching will not proceed.
6.4.2 Two of the most widely used halogenated
hydrocarbons are CF
4
and C
4
F
8
. The Matheson Gas
Book (79) says this about the chemical properties of
CF
4
, “...Carbon tetrafluoride is extremely stable,
reacting only slightly even at the temperature of a
carbon arc...”. About C
4
F
8
, “...Octafluorocyclobutane
is extremely stable. It is unreactive with other materials
under ordinary conditions. At high temperatures
(600°C), it dissociates to form carbon and carbon
tetrafluoride and some toxic compounds...”
6.5 Hydrides
6.5.1 References (1, 2, 3, 4, 5, 64, 65, 66) are cited
which indicate these gases do not react with silicon
under the prescribed conditions.
6.6 Halogens, other than fluorine
6.6.1 The halogens represent a special class of
materials. Molecular or atomic chlorine, iodine or
bromine have been reported not to etch silicon at
temperatures below 200° C or without the assistance of
a plasma (19, 26, 50, 55, 56, 63). Schwartz (74) makes
the comment. “Although Cl
2
chemisorbs spontaneously
on silicon, (in a plasma) at low temperatures the
reaction between Si and Cl does not proceed to the
formation of SiCl
4
. That is, silicon does not etch in Cl
2
except at elevated temperature.” He goes on to note:
“...except in the case of highly doped silicon, etching
occurs at low temperatures in the reactive ion etching
system not because of the ease of Cl attachment but
because partially reacted silicon can be sputtered
readily.” Schwartz then reports the observation that
there was no perceptible etching after thirty minutes in
a Cl
2
plasma when there was no cathode bias on the
wafer; in contrast to 4.5 microns of removal when the
bias was applied.
6.6.2 Flamm (55) published in 1981 and titled “The
Design of Plasma Etchants”, states, ”...Chlorine
containing halocarbon feed gases are frequently used
instead of Cl
2
because they are not hazardous (before
reaction), are noncorrosive, and they decompose in the
plasma to generate radicals (in addition to chlorine
atoms) which induce some desired effects.” This
statement applies to a plasma based system; Flamm is
pointing out the difficulty of etching silicon with just
Cl
2
or Br
2
in a plasma. He points out that Wang and
Maydan (from a private communication in 1979) have
etched silicon in a pure Cl
2
plasma at pressures below
0.05 torr.
6.6.3 Flamm goes on to say that, “...It seems that
chlorine and bromine atoms do not spontaneously etch
SiO
2
or undoped single crystal silicon but do etch some
forms of polycrystalline silicon and heavily n-doped
single crystal and polycrystalline silicon...”
6.6.4 A conclusion here is that even with a plasma Cl
2
and Br
2
require very specialized conditions to etch
silicon.
6.7 Halides, Other Than Fluorides
6.7.1 Gaseous mixtures of HCl, HBr or HI will not etch
silicon at the prescribed conditions (1, 2, 3, 4, 5, 62).
However, mild pitting of the silicon may be observed in
HBr or HI under aqueous conditions (5). In laboratory
testing (10) no reaction of the silicon was noted with
50% aqueous HBr after 45 days immersion.
6.7.2 One reason for the lack of reactivity of silicon
with chlorine, bromine or iodine halides is the low
vapor pressure of the reaction product. SiCl
4
has the
lowest boiling point at 57° C; SiBr
4
boils at 154° C and
SiI
4
at 286° C. On the other hand, the boiling point of
SiF
4
is –86° C. Practical advantages of this low vapor
pressure and non-reactivity were demonstrated by
Texas Instruments in their patents (24, 25) on RIE
etching; controlled additions of HBr, HI, BCl
3
, among
other gases, led to passivating films of SiBr
4
, etc. on the
silicon side walls and allowed vertical wall etching to
be achieved.
6.7.3 The moisture content of the gas stream can be a
critical factor in the corrosion resistance of most
materials. Typical stainless steel passivity is strongly
dependent upon the amount of H
2
O present, declining
sharply as water content increases above 1 ppm in the
presence of halogens or halides (6, 9, 10, 72). Paciej
(72) reported that 0.1 ppm moisture in HCl does not
attack 316L stainless, while 200 ppm moisture in HCl
can significantly corrode the surface of this material.
Fine (9) points out similar behavior for 316L in HBr at
0.5 ppm and 100 ppm moisture over the course of 10
days.
6.7.4 However, silicon corrosion resistance is virtually
unaffected by moisture content up to and including
aqueous solutions. The “RCA clean” (67) and
conventional wet etches (60, 68, 69) are obvious
examples of this.
6.7.5 Silicon based halides, such as SiCl
4
, SiH
2
Cl
2
, etc.
have not been reported to react with silicon under the
prescribed conditions. (55)
SEMI F79-0703 © SEMI 2003 3
6.8 Fluorine and Fluorides
6.8.1 It is documented that fluorine (F
2
) and atomic
fluorine, F, will attack silicon at room temperature (19,
22, 26, 42, 49).
6.8.2 Chen (49) used fluorine gas (F
2
) at room
temperature on freshly cleaned silicon; he measured an
etch rate of about 100A° /min at 2.3 torr and 5 sccm of
F
2
flowing. He also noted a two to three hour
incubation period before the etching started; the
authors speculate about the causes for this incubation
period.
6.8.3 Flamm (42) and Vasile (22) authored several
papers detailing the mechanisms involved in atomic
fluorine reacting with silicon. Flamm’s data produced
the following relation for the etch rate of silicon by
atomic fluorine.
6.8.4 R(Si) = 2.91*10
-12
*n
F
T
1/2
e-
E
etch
/kT
where E
etch
is given as 0.108 eV. At n
F
= 2.9*10
15
atoms/cc , the
etch rate is reported to be about 3,000 A° /min at 25° C.
6.8.5 Flamm used an RF discharge of at least 3.8 watts
to produce F from F
2
. Vasile used thermal dissociation
of F
2
to produce F; at 800° C he measured 47%
dissociation; below 650° C , negligible.
6.8.6 Flamm (55) also reports a similar relation for the
etch rate of SiO
2
by atomic fluorine: R(SiO
2
) =
6.14*10
-13
*n
F
T
1/2
e-
E
etch
/kT
where E
etch
is given as
0.163 eV.
6.8.7 Flamm (55) states “...Fluorine atoms react
spontaneously with all forms of silicon, SiO
2
and silicon
nitride to form volatile products.”
6.8.8 Ibbotson (19) examined potential etching
processes which required no external energy sources.
“Silicon is rapidly etched by the gas-phase halogen
fluorides ClF
3
, BrF
3
, BrF
5
, and IF
5
, in analogy to XeF
2
etching silicon..... By contrast, ClF and Groups III and
V fluorides such as NF
3
, BF
3
, PF
3
and PF
5
do not
spontaneously etch either Si or SiO
2
under the same
experimental conditions.”
6.8.9 Winters (21) and Ibbotson (20) report that XeF
2
does not etch SiO
2
. Regardless, silicon, with or without
additional oxide, is not recommended for use with F
2
,
XeF
2
, ClF
3
and the other fluorine containing inter-
halogens which may spontaneously decompose to
atomic fluorine at temperatures below 200° C.
6.8.10 As quoted from Ibbotson’s article above,
excluded from this list of reactive gases are compounds
such as NF
3
, BF
3
, PF
3
and PF
5
which have been
reported not to etch Si or SiO
2
without the assistance of
a plasma or RF source (19) or require temperatures
above 200° C (1, 35, 37, 38).
6.8.11 The ionized fluorine atom, F
-
, is reported to not
react with silicon under the prescribed conditions (1, 2,
3, 4, 5, 19, 64, 65, 66). Further evidence of this is the
fact that adding H
2
to fluorine containing plasmas
reduces the etch rate (55). The assumed mechanism is
the removal of F atoms as active species.
6.8.12 Other fluoride compounds, such as SF
6
, which
do not spontaneously decompose to atomic fluorine
below 200° C are not a source of reaction (33, 34, 52,
53, 55).
6.8.13 WF
6
is a special case. Numerous articles (41
,43, 44, 45, 46, 47) have been published on the CVD of
tungsten films from WF
6
. Two primary mechanisms
are cited; the reduction of WF
6
by silicon and the
thermal decomposition in the presence of H
2
. As stated
in Yarmoff’s paper (43):
“The dissociative chemisorption of WF
6
on Si (111)
was found to be complete, even at room temperature.
The reaction is self-poisoning at room temperature,
however as the fluorine liberated from WF
6
ties up the
active Si sites responsible for the dissociation.”
6.8.14 Tsao (41) reports that at 410° C only 200A° of
W will deposit from WF
6
onto single crystal silicon
which has an oxide thickness between 5 and 15 A° .
C.A. van der Jeugel (46) describes the effect of doping
levels on the self-limiting growth of tungsten films. It
is documented (44, 45, 47) that WF
6
will not deposit on
SiO
2
without the addition of hydrogen or some other
initiation mechanism.
6.8.15 Using the selective deposition characteristic of
WF
6
in a device, Fleming, et.al. have patented (48) a
technique for producing wear resistant coatings of
tungsten. Their invention requires a “clean” surface
and a temperature higher than 200° C, preferably around
450° C; exposure to WF
6
then results in a self-
terminating film of 5–50 nm.
6.8.15.1 Zdunek (80) reported on electrochemical
based work which indicated that unprotected silicon
surfaces exhibited “no degradation” after exposure to
WF
6
for 5 days at 80° C. A similar result was observed
for Cl
2
gas.
6.8.16 As a point of interest, metals such as stainless
steel, aluminum and nickel develop corrosion resistant
coatings of the respective fluoride when exposed to
atomic fluorine during the proper passivation procedure
(14, 23, 27, 28, 29, 30, 31, 32, 35, 39, 40).
6.8.17 Titanium, molybdenum, tungsten, brass and
columbium have been reported to be unacceptable
when exposed to gases such as ClF
3
, which decompose
to atomic fluorine (27, 32).