ML19322D511
| ML19322D511 | |
| Person / Time | |
|---|---|
| Site: | Crane  |
| Issue date: | 10/03/1979 |
| From: | Kastenberg W Advisory Committee on Reactor Safeguards |
| To: | |
| References | |
| ACRS-SM-0148, ACRS-SM-148, NUDOCS 8002140117 | |
| Download: ML19322D511 (21) | |
Text
'T 14c/&53%
?0/2 +/*/7:
October 3, 1979 TMI-2 HYDROGDJ GDJERATION: LICENSING IMPLICATIONS M4D DESIGN BASIS ACCIDEN'IS W. E. Kastenberg, ACRS Senior Fellow 1.
Introduction As a result of the accident at the tree Mile Island - Unit 2 (mI-2) reactor, the mI-2 Lessons Learned Task Force has made three "short term" reccanendations regarding post-accident hydrogen control systems for FWR and BWR containment (1):
a)
Dedicated Penetrations for External steccxnbiner or Post-Accident External Purge System, b)
Inerting BWR Containments, and c)
Capability to Install Hydrogen Recombiner at each UG.
Item (c) is a minority view with the majority opinion being that such consideration should be part of the long term reconsideration of the design basis for combustible gas contrcl systems. Wese recommendations are a result of the production of quantities of hydrogen gas in excess of the amounts required by NRC Regulations to be considered in the design and accident analysis of nuclear power plants. As stated in the report, "We Task Force is continuing to study whether the hydrogen design basis needs to be changed."
It is the intent of this note to discuss the hydrogen design basis, the potential implications of mI-2 h drogen generation on design basis i
accidents in general, and implications for future licensing and safety analysis (in particular, ice condenser plants).
8002140 117
cT, 2.
Brief History Following a toss of Coolant Accident, hydrogen can be generated by a) metal water reactions, particularly between the zirconitn I
fuel cladding and the reactor coolant (zirconium is oxidi::ed),
b) radiolytic decomposition of post-accident emergency cooling solutions (oxygen is also released in this process),
c) corrosion of metals by solutions used for emergency cooling or containment sprays.
i Generation by the metal-water reaction (M-W) is on the time scale of several minutes to a -few hours, while radiolytic decomposition time scales are hours to days, and corrosion time scales are even longer.
In the past major consideration has been given to radiolytic deemposi-tion because of the coevolution of oxygen and because the amount of H reaction was considered in the design basis is minimal. Consideration has been given to a) containment purging, b) use of recombiners and c) inerting NR Mark I and II containments, as a means of controlling hydrogen generation.
Discussions dating back to 1969, however, considered the possibility that with total ECCS failure, 25-50% of the clad in the core could react with water to generate hydrogen. In 1971, Safety Guide 7 was issued which required consideration of a 5% metal-water reaction (5% of the clad in the core) m LOCA and containment analysis. In 1974, an attempt was made
to have Safety Guide 7 conform to the ECCS Interim Acceptance Criteria which called for a 1% metal water reaction. Regulatory Guide 1.7, which evolved from Safety Guide 7 required either a 1% metal-water reaction or B
5 times the calculated value, whichever was higher.
The current design basis for combustible gas control is given in 10 CFR Regulatory Guide 1.7 - Revision 2, describes methods acceptable EC.44.
to the NRC Staff for implementing 5 50.44. With respect to metal-vater reactions, the extent and evolution time of initial core metal-water reaction hydrogen production from the cladding surrounding the fuel is
" Hydrogen production is 5 times the extent of the maximum calculated under 10 CFR Part 50, 5 50.46, or that amount that would be evolved from a core-wide average depth of reaction into the original cladding of 0.00023 in, whichever is greater, in 2 minutes."
Paragraph 50.46 specifies acceptance criteria for ECCS and along with Appendix K specifies an acceptable method of computing the hydrogen generation. Se value of 0.00023 inches corresponds to "one percent Since the reaction of the mass" for current designs with thin cladding.
is a surface phenomenon, the Staff believes that a strict 1% figure W e Staff would unnecessarily penalize reactors with thicker cladding.
further believes that a minimun metal water reaction (1% or 5 times that calculated in accordance with 5 50.46) *provides an appropriate and prudent safety margin against unpredicted events during the course i
i of accidents."
f l
l l
Since conformance with 5 50.46 ensures a maxims of 1% for the metal-water reaction (when considering frCS performance) one of two situations arises when considering combustible gas control:
I a) the calculated value is so low (as is 5 times it) that the design basis is 1% or b) the calculated value is just below 1% so that the design basis approaches 5%.
3.
TMI-2 Metal-Water Reactions Preliminary estimates of the hydrogen generated by the metal-water reaction at TMI-2 vary between:
a) 40% of the cladding wall thickness uniformly oxidized throughout the core (2), or more probably, b) 40% of the fueled region fully oxidized (2).
c)
A maximun of 48.3% and a minimum of 40.6% (3).
It was first estimated (4) that between 25% and 30% of all Zircaloy in the core reacted in the first 3 hours3.472222e-5 days <br />8.333333e-4 hours <br />4.960317e-6 weeks <br />1.1415e-6 months <br /> to produce hydrogen with j
the remaining reaction taking place between 3 and 9 hours1.041667e-4 days <br />0.0025 hours <br />1.488095e-5 weeks <br />3.4245e-6 months <br />, and later work (5) estimated 35%, with bounding estimates of not less that 20%
nor more than 60%.
Preliminary analysis (2) also indicates that at most, only 15% of 1
the total hydrogen generated evolved frm radiolysis. We figure is 1
)
l probably closer to 0% (3).
I 4.
Hvdrogen Generation and Design Basis Accidents W e design basis for ECCS acceptance, limits the metal-water reaction to less than 1%.
Similarly the design basis for combustable gas control
. lculated varies between a minimum of 1% and a maxim s of 5 times th W e % I-2 metal-water reaction (which cannot be greater than 5%).
d
- Clearly, metal water reaction involved over 40% of the Zircaloy cla.
We pressure spike E
the design basis in each instance was exceeded.
i g of
[
at 10 hours1.157407e-4 days <br />0.00278 hours <br />1.653439e-5 weeks <br />3.805e-6 months <br /> (28 psig over ten minutes), attributed to the burn n f 60 225 lb - moles of hydrogen, was however below the design basis o psig for the containment.
Guide 1.7 Part of the rationale for including the margin in Regulatory i
t considerations, was to separate the ECCS considerations from conta nmen f ECCS degrada-and to provide margin for hydrogen control in the event oh We periods of time in which the ECCS were turned off by t e tion.
(or failure).
1MI-2 operators essentially represented total degradation i
s a cause In the Reactor Safety Study (WASH-1400), hydrogen combust on a d BWR sttdies (5).
of containment failure was considered for the IHR ar i h a sub-atmospheric (The PWR sttdied was the Surry Reactor, at 788 MWe, w t We BWR studied was the Peach Bottom Reactor, Unit II, dry containment.
It es estimated that in a a 1065 MWe plant with a MARK-I containment).
i conium would core meltdown accident (failure of ECCS), 75 + 25% of the z r f hydrogen.
react with the water in 45 minutes yielding 600 lb - moles o ilure of contaiment Although a hydrogen explosion was ruled unlikely, fa j
i with a due to over-pressure caused by hydrogen burning or in con unct on Such events wre contributors to Category steam explosion was considered.
2 releases for both the PWR and BWR studied.
%ese release categories, in terms of fission product release, exceeds the TID-14844 values (100% r.oble gases, 50% halogens and 1% of the solid fission products), exceed the source terms given to Reg. Guide 1.4 and 1.5 (for PWRs and BWRs), and result in site boundary doses which may B
exceed 10 CFR Part 100.
Following the methodology of WASH-1400, Chan (6) exmined potential failure modes of alternate containments (PWR-Ice Condenser and BW He found that the ice con-Mark III) of the vapor suppression type.
denser containment can be failed in several ways under core melt Chan estimates a pressure conditions, including hydrogen burning.
with breach-rise of 50 psi in the contairraent due to hydrogen burning If the amount of metal-water reaction is kept below ing of containment.
For the BWR Mark III case, 25%, failure was estimated to be precluded.
Chan estimated that a 100% metal-water reaction followed by hydrogen burning yields a peak pressure of 156 psia, well above the 40-50 psia failure pressure range.
Such postulated accidents (e.g. core melt followed by contairsnent failure) are considered beyond the design basis, and as such are generally not considered in the licensing process.
The physical phenomena following a hypothesized core-meltdown accident ld were recently examined and the quantities of fission products that wou be expected to be released from containment were determined for the Of particular interest is the potential failure of Sequoyah EWR (7).
l 1
the containment (Sequoyah is an ice condenser) due to hy!rogen burning.
We Sequoyah containment is designed to accommodate an internal pressure of 10.8 psig. It is believed that the nominal failure pressure is 27 1 3 psig (4213 psia)(7).
We potential for contairn at rupture due to hydrogen burning depends on composition of the atmosphere, availability of an ignition source and an incremental pressure rise associated with the burning. For the Sequoyah ice condenser PWR, it was found that conditions favorable to hydrogen burning can be met with a well-mixed contairunent atmosphere (e.g., with the Air Return Fans operating) and if the path from the core to the con-tainment atmosphere is short (e.g. for a hot-leg break where the hydrogen may be above the spontaneous ignition temperature). Analysis indicates that failure "would be a virtual certainty", assuming hydrogen burning, because of the low design pressure.
In addition to these results, fission product release for various contain-ment failures was considered. Using the methodology given in Appendix VII of WASH-1400, LOCA sequences which include hydrogen burning fall into Release Categories 2-5.
For the sequence (S HF-1) falling into Category 2, the 2
following releases wre obtained:
Xe-Kr 100%
Ru 4.2%
I-Br 13%
La 0.7%
I Cs-Rb 57%
Te 49%
4 Ba-Sr 6.8%
I
is a small pipe break (1/2 to 2 in) with failure of the (Note: S HF-1) l ECS recirculation system and failure of the containment spray recircu a-2 tion system.)
5 i
These fission product releases are also in excess of the design bas s source terms.
5.
Potential Solutions _
Tne discussion above raises several questions concerning hydrogen gene d future licensing tion and control, currently used design basis accidents an Of particular concern are ice condenser contairsnents, which becau reviews.
burning.
of their low pressure capability, may be vulnerable to hydrogen il In addition to hydrogen recombiners, there are three (or more) potent a methods for dealing with the ice condenser containments:
inerting the containment, a) provide aux 111ary means to suppress hydrogen b) burning, and employ post-accident filtration (vented containment).
c)
I Inerted Containment _
5.1 i
Me requirement of inerting BWR-MARK I containments was a means o In principle, the inerting requirement with early versions of 10 CFR 50.44.
i for hydrogen could be placed on ice condenser plants if the design bas s Bere are several negative aspects of inerting ice generation were changed.
condenser containments which relate to access.
. At the D.C. Cook plant (the only operating EWR with an ice condenser containment), it is req d red to monitor ice build up on the vent doors (see Figure 1) on a weekly basis. Wis ence a week entry has In addition, been made at the suggestion of the resident inspector.
I there is more equipaent requiring routine maintenance inside a FWR containment than inside an inerted BWR Mark-I containment.
3 ft, BWR-Mark I containment takes 1 day to inert and 1 A 160,000 day to deinert, at a cost of several thousand dollars for the liquid An ice condenser containment, with a free voltane of nitrogen.
3 should take a somewhat longer time for inerting and 1,192,000 ft deinerting, and at a larger cost.
5.2 Auxillary Means to Suppress Burning te provision to suppress hydrogen fires by auxilary means could take We most conventional approach for combustible gas control several forms.
is the use of recombiners, now in use commercially and at several power Because the state-of-the-art is such that long periods of time reactors.
are required for extensive recombination (several days), recombiners are best For example, the suited for control of hydrogen generated by radiolysis.
Sequoyah plant will have two electrical thermal hydrogen recanbiners based on a 1.5% zirconitza-water reaction and an 8 day period to reach combustable Alternatively, controlled burning of hydrogen as it is produced is limits.
itile large dry PWR containments may take the pressure also a possibility.
i i
. rise associated with this option, it is questionable for an ice con-denser, especially if the hydrogen is generated very rapidly (on the order of minutes).
We physics of hydrogen flammability can also be used for control.
In particular, the presence of a taird vapor or gas will influence the flamability limits of hydrogen in air and oxygen atmospheres.
Flammability of two vapors can be suppressed by:
a) a suitable increase in the amount of either constituent, b) the addition of a suitable amount of inert substance or an oxygen scavenger, c) the addition of a flammable substance in sufficient amounts to exceed the higher flan-mability limit of tb resultant mixture.
Figure 2 shows the effects of steam on the flanmability limits of hydrogen /
air mixtures. When approximately 60% water vapor is present (80 C), the flammibility limits coincide at 10% hydrogen /30% air. Hence flammability is suppressed for all hydrogen concentrations up to 40% providing there is 60% water vapor present.
Other tertiary mixtures (with hydrogen and air) might include nitrogen, carbon-dioxide, methane and hydrazine. Hydrazine is of interest because it is an oxygen scavenger and has been employed in contairunent spray systems as an iodine getter (8).
(In the Sequoych containment, the ice will contain sodiun tetraborate to enhance iodine adsorption).
1
\\ In this latter a@ roach, one could make use of the containment spray
~
One alternative would be systems to effectively " inert" the contalment.
ih to increase the temperature of the contaiment slightly and " fog" it w t We objective is a mixture of steam / air / hydrogen above the spray system.
E ii essure below the flammability limits but at the same time mainta n ng a pr A second alternative would be a chemical the contaiment design pressure.
As employed now, hydrazine additive, such as hydrazine in the spray system.
is added to a boric acid / water solution by a metering pump before it enters We quantity of hydrazine used (100-200 gallons) is the contaiment.
large in comparison to Iodine - 131 present because of radiation depleti For either alternative, the contaiment spray system would of hydrazine.
be turned on imediately following an accident.
If such an approach as described above is feasible, there is one item of In WASH-1400, a major risk contributor for concern; loss of all AC power.
the PWR setzlied was a loss of offsite AC power and failure to recover l
either onsite or offsite AC power within three hours, (followed by fai ure It appears, however, that the emergency of the feedwater delivery system).
AC system design for the Sequoyah plant includes an additional level of While this redundancy may eliminate redundancy over that of Surry (7).
)
the event as a major sequee e initiator, its role (loss of all AC power in other sequences is not clear.
l It should also be noted that other methods such as the use o i
elements to form hydrides, molecular diffusion through a diffusive barr er,
. chemical fixation, regencrative liquefaction and utilization of antide-flagration agents (e.g. bromotrifluoromethane) may prove useful for this Sufficient information, such as effectiveness in radiation application.
environments, needs to be established.
5.3 Post Accident Filtration he provision of a post accident filtration system (PAFS) or vented contalment as a means of coping with hydrogen generation stems from the general attempt to improve contalment effectiveness under core melt conditions. The objective, as described by Gossett et al (9), is to provide an external filter through which the air / steam mixture in the containment could be vented to the atmosphere, thus It also preventing containment rupture due to overpressurization.
has the secondary function of removing radioactivity from the contain-In the ment, even in situations where the contaiment is leaking.
original work by Gossett, a PAFS for the BWR-Mark I containment was designed and analyzed. Preliminary results for a PWR dry containment were also presented. A unique feature of these designs is the inclusion of a hydrogen ignition chamber so that unintentional burning or detona-tion in the filter system is precluded.
When considering a PAFS for an ice condenser containment in conjunction with hydrogen generation, other contalment failure modes must be considered. Were are several potential failure moles of interest:
, a) hydrogen burning b) steam explosions c) containment leakage (failure to isolate) d)
core debris fragmentation e) overpressurization due to noncondensibles, steam or both.
Because of the low design pressure of the ice condenser, failure due to a) and e) are the most probable for a broad spectra of accidents, even if the contalment ESFS operate (7). Hence, the primary function of the PAFS is an ice condenser would be to vent the contalment before substantial pressure buildup due to barning and/or the presence of steam and other noncondensible gas.
An imoortant consideration in the design of a PAFS for an ice condenser contalment is the interaction between it and the other contaiment ESF systems (containment spray, air return fan system and the ice itself). In particular, the air return fan system which is actuated i
on high containment pressure, is delayed for 10 minutes following initiation of the ECCS. We fans, when operating, reduce contairrnent pressure. W e 10 minute time delay is intended to provide an increased back pressure during core reflood. Pumps which draw suction from the sump may cavitate if the water were to flash due to a sudden drop in l
l pressure. Mien the fans are not operating, hydrogen would tend to l
accumulate in the lower compartment, (368,000 cubic feet). With the
fans on, the hydrogen would be distributed throughout the containment.
After long periods might accunulate in the upper compartment (should the fans stop working).
E j
Given the compartmentalization of the contairunent, it is not clear Flow rates in the where the entry points for the PAFS should be.
PAFS should be compatible with the air return fan system and the natural driving force of the air / steam mixture through the ice The operating mode is also dependent on the other ESF compartment.
For large dry containments, the PAFS would be activated functions.
upon high pressures being reached (50-60 (isolation valves opened)
For the low pressure ice condenser, the operator has to be psig).
prepared to activate the system very soon after initiation, but not Inadvertant operation of so soon that other ESFs are compromised.
PAFS may also be a problem.
Last but not least, is the public's attitude toward controlled The deliberate release of some species, venting of the noble gasses.
however small their radiobiologic effect might be, might not be Consideration should be given, in acceptable in the public view.
terms of cryogenically cooling the charcoal filters, to controlling l
noble gas release.
l
s 1
& i 6.
Reconnendations mile it is too premature to recomend any (or all) of the potential For the g
options discussed above, the following might be appropriate.
short term, consideration should be given to inerting ice condenser Wis would involve lengthening the periods between containments.
entry (perhaps on a monthly basis) and/or providing air-paes to those entering while the containment is inert. Figures 3 and 4 indicate that if the oxygen is reduced to below 4.9% (from 21%),hWrogen ignition is precluded.
For the long term, research should be carried out on the use of chemical additives in conjunction with the contaiment spray system l
We use of containment and in the use of filtered vented containment.
Vented spray has the advantage that the system is already in place.
contaiment has the advantage that it can potentially cope with other sources of containment failure (e.g. overpressure due to steam explosion and/or non-condensible gases).
Sandia Laboratory has recently initiated a program, under the sponsorship of NRC, to investigate vent-filtered containment conceptual designs for It is recommended that NRC direct Sandia to exmine light water reactors.
conceptual designs for ice condenser plants at the earliest time so that a decision regarding their possible use can be made en s ;'.nely casis.
. 7.
Concluding Remarks te available information regarding mI-2 indicates that the design basis for hydrogen generation ws greatly exceeded (40% metal /wter reaction observed), and the pressure spike of 28 psig at 10 hours1.157407e-4 days <br />0.00278 hours <br />1.653439e-5 weeks <br />3.805e-6 months <br /> in the contairunent as due to hydrogen burning. In assessing the implications of m 1-2 with respect to hydrogen, particular concern focuses on PWR-Ice Condenser Containments Wich have a low pressure contairunent design (10-12 psig).
Preliminary analysis of ice condenser containment failure modes indicates that for a broad spectrum of accidents, beyond the design basis, failure is almost certain from hydrogen burning or overpressuri-zation due to steam and/or noncondensibles. As presented in Sections 5 and 6, there are several potential options for dealing with hydrogen generation beyond the design basis. @ ese include inerting, vent-filtered containment and use of the containment spray system.
Rese considerations raise several important questions:
1)
Wat should the design basis for hydrogen generation be? Should hydrogen control be predicated on metal-wter reactions (short-term) or on radiolysis (longer i
l term).
I 2)
Should accidents Wich currently lie beyond the design basis be considered in future reviews? Are accidents i
dich terminate with a disruptedJcore (not necessarily a molten one) be included in the design basis?
i
s
- 17 Considering the extent of metal-water reaction possible 3)
~
(greater than 404), with failed clad and disrupted fuel l
(not necessarily molten), are currently acceptable source terms (fission products) still adequate for site evalua-E tion?
In view of (3) above, are current or planned methods of 4) hydrogen control adequate? Will such processes as recem-bination, purging or venting provide larger releases of What radioactivity than previously thought possible?
are the possible interactions between fission product release and hydrogen control?
References:
71I-2 Lessons Learned Task Force Status Report and Short-Term 1.
Recommendations, NUREG-0578, July 1979.
Memo R.O. Meyer to R.J. Mattson, April 13, 1979.
2.
Memo W.R. Butler to R.L. Tedesco, April 25, 1979.
3.
4.
Memo M.L. Picklesimer to File, June 20, 1979.
R.K. Cole, Jr., " Generation of Hydrogen During the First tree Hours of the W ree Mile Island Accident," NUREG/CR-0913 SAND-79-1357, 5.
July 1979.
(See also WASH-1400 - Appendix VIII, Sections 2.2.7 and 3.2.7.
6.
Appendix D to Appendix VIII).
Chan, C.K., "On the Failure Modes of Alternate Containment Designs following Postulated Core Meltdown, UCIA-ENG-7661, June 1976.
7.
Reactor Safety Study Methodology S.V. Asselin et. al, Draf t Report:
Applications Program, Sequoyah #1, PWR Power Plants, 1979.
8.
Tam, P. "We Effectiveness of Additives in Contairrnent Spray 9.
17, 1979.
Systems" Letter to D. Moeller, August Gossett, B. et. al, " Post Accident Filtration as a Means of Improving Containment Effectiveness", UCIA-ENG-7775, Dec.1977.
10
t h
STEAM GENER ATOR
_I.'
gl ENCLOSURE M f~=n t !
u.
N,*,'k (RAst j
/- " "'
Kh' 108 98C E T
i p--
-)
.u eme e m
m,,
~
1,
. su
-e -,
C.Mts I
1 n
\\
m.
.m-.(g (Deurke l
ClWtseestaf.w E, BaMU t
I
",5 M.
I
[
f'
.uff er,g mu e=.s I
i 6
A 1-w-x. o. a l
e,
i r.-
i l
si., cia I
vaa
.. uia. --
- --i te=>
y i
'._;.l-l l
i
/N g.
I
~
!.. h I c,...,,,
5' i
p.
.\\
/_
l.,,,
__s.
..id=5=';
,"-'fL
-yama. j~~__
---i li
! ', 1 O.,
6 4
~
i <,
- I l
e
.M tp.
. s s s;
,1 f--<*x*'
m' i
- e, 3l y
i 0"",
1 s
/
./ :i T --- '
4
- y.t=
t at
,u,
/
h LL L;
I
. e..,a.<.,,=
1 i,
-.+
.. m, s
i.....
s i.
q: s W3 N.ha.3 Q1
.o.
. c.
.3 i;
ii.'
FAN Ai ACCUMULATOR g
)I. }
Rooms rf.,-
p Count s.stsa.
..am.a...e.nt.att.s p
.t FIGURE 1 --- Sectional Elevation DD -- D. C. COOK unit h'o.1 Showing Ice Condenser Containment Volume Boundaries
i l
Ny 4 F le -u4.4$
y[ W dyp e4 fua - i c C /l b/ 4 f D - J C
- TVS, h / if 5 h i
1 i
i 1
1 1M 0 FLAMMABILITY LIMITS i
75 F O psig
---0---300 F O psig
- -e
- 300 F 100 psig
\\
2
\\
80 20 t
40
~
60
[
ASSUMED g*
,A t'
DETONATION 1 *\\
O Y
LIMIT,S
/
v 60 g 40 W=r{f'Q'6
- FLAM5tABILITY LIMITS
/ \\
/h N
,,g.r-100 100 80 60 40 20 0
0 HYDROGEN, v/o PURE H2 FIGURE 2 - Flamer. ability Limits of Hydrogen Air Steam Mirtures G
l
l j,g%m%We. nzu.rm _._
Lturrs or rtum-rm OF GASES red YAPOSS 20
'!%e determinations were are shownin 6gure 7.
In8uence of Temperature.-In a 3ke.elooed made in a tube 6 feet in length and 2 inches in bulb the limits were 9.6 and 90.9, percent at 15' diameter, with upward propagation of Same at C. and 9.1 and 94 percent at 300 C. (SS7).
ured p
a t n (1 htmosp era ETDROGEN DI QTlWL ATWOSPMIRES it may,be calculated that no mixture of hydro-j gen, nitrogen, and air at atmospherie pressure All Atmospheres of Orygen and Nitrogen.-
and temperature can peopagate dame if n The limits of hydrogen in various mixtures of oxy;cn and nitrogen have been determined at contains less than 4.9 percent oxyFen U671 ful 600 mm. and lower pressures, with downward For some purposes the results are more use 5
propagation of flame in a tube 3 cm. in diameterwhen azpressed (62) as in 6gure 8.
(&s also Ammonia Contact Gas.)
This figure shows, for example, that a mixt ute containing 20 percent H,,6 percent O2, and 4 (661Atmospheres of CompositionBetween Air and percent %. is Sammable; but if 2 percent of the Pure Oxygen.-With downward propacetion of
'f!nme in a Bunte burette, the lower limit fell oxygen were replaced by nitrogen the mixture gradually from 9.45 percent hydrogen in air to would not be flammable but would become so by admixture with a suitable amount of air.
O.l.*> percent in nearly pure oxygen. The i
In figurc 8, "imponible mixtures" carmot be higher limit rose from 65 percent hydrogen in air to 51 percent in a 40-percent oxygen mix.
produced by mixing air, nitropen, and hyd ture. 50 percent in n 56. percent oxygen mixture, azul 91.0 percent in nearly pure oxygen (323J.
corresponding section on methane h'mits in in a mi. ture of equal volumes of oxrgen and mixtures of air and nitavgen (pp. 44 to 461 s
nitrogen. 91.35 percent hydrogen inflamed at The limits with downward propagation of Same in the acme series of mixtures bare been 53P C. (21Cf.
Atmospheres of Air and Nitrogen (Air De-determined in a clowd tube 5 cm in diameter ficient in Oxygen).-The. limits of hydrogen in and 65 cm. in length. The lower limits are 5 ull mixtures of air and nitrogen. or air from I
f which part of the oxygen has been removed, to G percent greater and the higher -1 to :-10 01tcEN Di osuc!NAt. ATWOSFMERE.FERCIFT so to is 24 12 to a
s 4
e a
j e
i i
i i
f f
i i
i m
l
~
N l
\\
N t
I
\\
so aox%
IQ J
PN
~
I N
o o
o o
o o
o en cARaoM DeoEmE on ADemONAL MrTROGEN Di OBUGDiAL ATNOSP o
FIGljRE 3 -- Ltmits or riamm.buity et sydrosen in Arr and carboa Deside oc xitros n.
g t
=.
l
. n - s. ~ e - "
bk%
- IGbi, hww.MMWh, hAA b4. @CLD 21 LDEITS OT INDn*1 DUAL CASES AND VAPORS
=
_\\\\
~
\\
l
\\
s i
l
\\
I j
l Dies mesMe af hasmas hie enesteve I
]
e l
\\
it 1
E l
I 5
r s== = 6i *-
i 1
I t
l Capaw. t mm.as eiens.-. =m=. w,
l l
,, tenniaisi ass aw bydroge is emptmee per se il se D
(B 40 eb We dit 5a kil0 irrnacco.rsacsxt Tir. cat 8.-Relation Between Compositico and Flanmobility of.tiuures of Hydrogen. Occa n i
mixtures of air and carbon dioxide are shown in The percent greater than those of figure 7.
figure 7.
The determinations were made in a
- of the curve is at the same percentage The addition of 0.5 tube 6 feet in length and 2 inches in diameter,
" note with upward propagation of Same at atmos-of ndditional nitrogen.
percent of tin tetramethyl reduces the higher pheric pressure durm propagation (13J,167).
1 mit and retrnets the I' note" of the curve The limits with ownward propagation of considernbly (JIS). The limits for dowitward flame in the sarne series of mixtures have been propagation in a closed tube 2.2 cm. in diameter determined in a closed tube 5 cm. in diameter t
and 65 cm. in lengtb. Thelowerlimits are 5 to 6 have also been determined (J.ll).
Atmospheres of Air and Water Vapor.-The timits of hydrogen air mixtures standing over percent greater and the higher limits 1 to 4 less than those m figure 7.
The water in a 35CRe. spherical vessel, and ignited percent" nose" of the curve ia at 56__ percent carbon near the water surface, have been determined at dioxide in the atmosphere. The addition of temperatures. As the temperature 0.5 r>ercent tin tetramethyl reduces the higher various riu s. and consequenti,v the water. vapor content limit and retracts the " nose" of the curve alm. the lower lunit rises slowly, and the higher considerably (318). The limits with downward
- limit falls rap dly, as with other diluents. When propagation in closed tubes 2.2 and 1.6 mm. in
}i 60 reent o water vapor is present 86 C.)
diameter have also been determined (fl7,341).
h drogen 1[ the imits coincide at about 10 percent Some earlier observations (SJ) show, as mi ht be expected. a more rapid narrowing of he (Jcs).
Earlier experiments, made in a Bunte burette, limits in a Bunte burette. Others (1) may be abow similar effects but the range of flamma-mentioned, but they can hardly be s.ecepted without confirmation because they indacate bility is smaller (95).
Atmospheres of Air and Carbon Dioxide.-
several improbable conclusiona-for example, The limits of flammability of hydrogen in all ame -m-a
.