Determination of the Speed of Flames and the Temperature Distribution in a

Spherical Bomb from TimePressure Explosion Records

Bernard Lewis and Guenther von Elbe

Citation: The Journal of Chemical Physics 2, 283 (1934); doi: 10.1063/1.1749464

View online: http://dx.doi.org/10.1063/1.1749464
View Table of Contents: http://scitation.aip.org/content/aip/journal/jcp/2/5?ver=pdfcov
Published by the AIP Publishing

Articles you may be interested in

A spherical bomb
Phys. Teach. 14, 69 (1976); 10.1119/1.2339312

Experimental Determination of Rotational Temperatures and Concentrations of OH in Flames from

Emission Spectra
J. Chem. Phys. 23, 143 (1955); 10.1063/1.1740513

Determination of Transient Flame Temperatures

Rev. Sci. Instrum. 25, 640 (1954); 10.1063/1.1771149

Measurement of Temperature Distribution in a LowPressure Flat Flame

J. Chem. Phys. 22, 824 (1954); 10.1063/1.1740197

TIMEPRESSURE CHARACTERISTICS OF VARIOUS DIFFUSION AND MOLECULAR PUMPS

Rev. Sci. Instrum. 3, 309 (1932); 10.1063/1.1748940

This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
128.248.155.225 On: Sun, 23 Nov 2014 22:06:54
 MAY, 1934 JOURNAL OF CHEMICAL PHYSICS VOLUME 2

Determination of the Speed of Flames and the Temperature Distribution in a

Spherical Bomb from Time-Pressure Explosion Records 1
BERNARD LEWIS2 AND GUENTHER VON ELBE,3 U. S. Bureau of Mines
(Received February 2, 1934)

A method has been developed for determining the atures, etc., have been made and tabulated for explosions
velocity of flame relative to the mass movement of the of mixtures of ozone and oxygen. The speed of flame
gases, in a closed spherical bomb from an analysis of the increases from the center of the bomb to the wall. At the
time-pressure record of the explosion. The speed of the same time the pressure and temperature of the gas about
flame can be evaluated at any moment during its progress to be burned increases. The temperature gradient in the
from the center to the periphery of the bomb, as wen as gas from the center to the wan has been calculated when
the temperature existing in the unburned phase, the the combustion is complete. A complete diagrammatic
temperature immediately behind the flame front, the description is given for one explosion. It is shown that the
temperature gradient from the latter point to the center temperature gradient actually existing in the bomb does
of the bomb, and the pressure in the bomb at the same not affect the specific heat results obtained by the usual
moment. Given a certain fraction burned of the total method of calculating the final temperature from the
amount of gas, the volume occupied by the products can maximum pressure by means of the gas law. It is pointed
be determined for three conditions: (1) Before it has out that the practical coincidence of the pressure curve
expanded against the rest of the unburned gas; (2) after with the zero line in the first part of the time-pressure
it has expanded; and (3) when combustion is complete record is not due to a time lag between passage of the
and it has been compressed by subsequent burning of gas spark and ignition, but to the smal1 fraction of gas which
nearer the periphery. Calculations of flame speeds, temper- has burned during this time.

INTRODUCTION of the soap-bubble method are preserved, while

its disadvan tages are eliminated. Experimen tally,
A METHOD of measuring the velocity of
flames was successfully developed by
the method is both simple and applicable in
general. It will be apparent from the following
Stevens. 4 It consists of igniting a combustible
that a single explosion record provides a whole
gas mixture at the center of a soap bubble and
series of data on the explosive mixture under
photographing the spherical flame as it expands
consideration, from which can be determined
at constant pressure. The method is simple and
the flame speeds at different temperatures and
direct. It eliminates the serious disadvantages
pressures, the specific heats of the burned 5 and
of flame-speed measurements in tubes, but it
unburned gases6 separately, the temperature of
suffers from certain limitations. First, it ·cannot
the unburned gases, and the distribution of
be used for gases that attack or are affected by
temperature in the burned gases at any time
a soap film. Second, the gases are necessarily
during the explosion up to and including the
always contaminated with water vapor.
time the flame reaches the wall.
On the other hand, time-pressure diagrams
The experimental procedure has been de-
obtained from explosions in a closed spherical scribed in detail elsewhere. 7 Briefly, it consists
vessel can be analyzed successfully for the of measuring the increase in pressure during the
determination of flame speeds. The advantages explosion of a gas mixture ignited at the center
of a closed spherical bomb. The change of
1 Published by permission of the Director, U. S. Bureau
pressure with time was recorded on a rotating
of Mines. (Not subject to copyright.)
2 Physical chemist, Pittsburgh Experiment Station,
film by the movement of a light point resulting
U. S. Bureau of Mines, Pittsburgh, Pa.
3 Member of staff, Coal Research Laboratory, Carnegie • Lewis and von Elbe, J. Am. Chem. Soc. 55, 511 (1933).
Institute of Technology, Pittsburgh, Pa. 6 See following paper, p. 291.
4 F. W. Stevens, J. Am. Chem. Soc. 48,1896 (1926). 7 Lewis and von Elbe, J. Am. Chern. Soc. 55, 504 (1933).

283

This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
128.248.155.225 On: Sun, 23 Nov 2014 22:06:54
 284 B. LE\VIS AND G. VON ELBE

from the deflection of a specially constructed The relation between pressure and the amount
diaphragm manometer. of gas burned has been developed by Mache. 8 • 9
When a gas mixture is ignited at the center of In reproducing this equation we shall employ
the bomb the fiame spreads spherically and more familiar symbols relating to mols instead
concentrically toward the wall. Since the flame of unit masses, as used by this author. Let
movement is slow compared to the velocity of V = the total volume of the bomb.
sound, pressure equilibrium is considered to vb=the volume of the burned gas at pressure P.
prevail at all times. The gases burning initially Vi = the volume that the mass of gas enclosed in Vb
at the center of the vessel expand to a pressure, occupied before ignition at initial pressure Pi and initial
P, compressing thereby the unburned phase temperature T i •
ahead. The work done on the unburned phase p. = the final pressure (maximum) established when the
entire contents of the vessel have reacted.
raises its temperature from its initial to some Cb and Cu = the mean molar heat capacity of the burned
higher value. The burned phase loses an amount and unburned gases, respectively, at constant pressure
of thermal energy equivalent to the work done. between temperatures Tbi and Tb and Ti and T", respec-
As the flame progresses the pressure increases, tively, where T" is the temperature of the unburned phase
and Tb is the temperature bordering directly on the burned
further compressing the unburned phase as well
side of the reaction zone. Tb is identical with the tempera-
as the burned phase; in consequence the temper- ture the gases reach when allowed to burn at constant
atures of both phases are increased. pressure at temperature Tu. Likewise Tb. is the temperature
The principal object of this report is to reached when the gases burn at constant pressure at
temperature Ti.
determine the velocity of the flame relative to
1'b and 1'u = the ratios of heat capacities at constant
the moving unburned gas at any point in the pressure and constant volume of the gases in the burned
enclosed sphere or, in other words, the velocity and unburned phases, respectively, for the same tempera-
of propagation of the reaction .zone into the ture ranges as Cb and Cu.
unburned gas if the latter is considered to be nu and nb = the number of mols per unit mass of un-
burned and burned gases, respectively.
at rest. This is the true speed of the flame.
It is smaller than the flame velocity relative to In general, indices i (initial) and e (end) refer
the fixed coordinates of the bomb by an amount to the states before ignition and after complete
equal to the mass movement of the gases. combustion, while indices u (unburned) and b
(burned) refer to the unburned and burned
phases during the progress of the combustion
DEVELOPMENT OF THE METHOD
from the center to the wall of the bomb.
The first step toward solution of the problem Furthermore Mache introduces the relation
is to develop the relation between the amount of
(1)
gas burned up to a certain time and the pressure
established at that time. From this equation and where .K is a constant. The only assumption is
the experimental time-pressure record of the that mean molar heat capacities Cb and Cu are
explosion the amount of gas burned during any practically constant over a narrow range of
time interval can be calculated. This amount of temperature. For Cu the maximum range of
gas will be understood to occupy an annular temperature extends from Ti to the temperature
shell. If the time interval is the unit of time, of the unburned phase at its maximum com-
then the thickness of this shell just before pression, namely, Tue. Correspondingly, the
burning represents the distance the reaction maximum range of temperature for Cb extends
zone penetrates into the unburned gas in unit from Tbi to 1\" the latter of which is the temper-
time, that is, the desired flame velocity. The ature reached when combustion (at the end)
thickness of the shell can be calculated only takes place at constant pressure at temperature
Tue. Neither Cb nor Cu are exactly constant in
when the amount of gas contained in it, the
prevailing pressure, and the radius of the shell
8 L. Flamm and H. Mache. Wien. Ber. 126, 9 (1917).
and its temperature are known. The latter is • H. Mache, Die Physik der Verbrennungserscheinungen,
the temperature of the unburned phase. 1918.

This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
128.248.155.225 On: Sun, 23 Nov 2014 22:06:54
 DETERMINATION OF THE SPEED OF FLAMES 285

this temperature span (100° to 200°), but for all temperature. Its subsequent changes of state
practical purposes Eq. (1) appears to be fulfilled, with increasing pressure in the vessel are
as wiII be shown later. governed by the same adiabatic compression law,
Then if v;/ V is the fraction of the total gas namely,
burned at the time pressure P is established, Pi! 'Ybdvb = cons tan t. (2b)
However, the conditions existing when this
subsequent compression starts are different for
the various burned shells, considering that the
innermost shells, for example, burn practically
where R is the gas constant.
at pressure P. from an initial temperature close
The fraction v;/V must equal 1 when P=P e
to Ti while the shells near the wall burn practi-
and zero when P=Pi,. The first condition is met
cally at pressure P e from an initial temperature
by Eq. (2). The second condition allows the
close to Tue.
determination of K by Eq. (2) using pressure
In order to apply the adiabatic compression
and temperature data of an actual explosion
law to a shell after it has burned, its temperature,
experiment. One must demand that this value
T b : immediately after burning must be known.
of K agree reasonably well with the value of K
calculated by Eq. (1) using band spectra specific For this purpose Mache introduces Eq. (1).
Eqs. (I), (2a) and (2b) in conjunction with the
heats. Furthermore, the practical constancy of
condition of constancy of the total mass and
Cb and C" must be demonstrated by the agree-
volume, transform to Eq. (2) by a series of
ment in the values of K calculated by Eq. (1)
mathematical operations.
for different combinations of Tb and T", These
It can readily be shown that at the moment
comparisons wiII be made later in the application
maximum pressure is established the gas is
of the method to gaseous ozone explosions.
hotter in the center than it is near the wall. lo
T" is given by the adiabatic law:
An elementary mass of gas at the center burns
(3) by expanding at practically constant pressure P j,.
It is subsequently compressed to nearly its
Tb is given by the thermodynamic relation:
original volume by the combustion of the rest
Tb=(qp/(nb)CPTuTb)+T,., (4) of the gas in the vessel. The latter work of com-
pression exceeds the former work of expansion
where qp is the heat liberated per unit mass of because the compression of the elementary mass
gas at constant pressure and CPTu Tb is the mean takes place at a steadily increasing pressure
molar heat capacity of burned gas at constant from Pi. to Pe, while the expansion takes place
pressure between T u and T b. at the lowest pressure Pi. An elementary mass of
It may be desirable to show briefly how Eq. gas near the wall, on the other hand, is first
(2) was obtained. compressed under a pressure which rises from Pi
Consider the mass of gas at pressure P to P e and then expands (on combustion) to
enclosed in an elementary volume dv u in the approximately its original volume at the maxi-
form of an annular shell. Before the gas is burned mum pressure Pe. The latter work of expansion
it is compressed by the rising pressure, according is obviously larger than the former work of com-
to the adiabatic law, pression. Thus the elementary mass of gas near
plhudv u = constant =Pi,lhidvi" (2a) the wall loses some of its energy while the
elementary mass of gas in the center gains
where dVi is the shell volume occupied by the energy in excess of the chemical energy released
same mass of gas before ignition occurred. It within it.
appears from Eq. (2a) that the change of state The radius of any shell is calculated as follows.
of the unburned phase is unambiguously given According to the gas law, the total number of
by the initial 'conditions for any pressure P. 10 Experimental proof of a temperature gradient was
When the mass of gas in the element considered first given by Hopkinson. Proc. Roy. Soc. A77, 387
burns, it suddenly changes its volume and (1906).

This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
128.248.155.225 On: Sun, 23 Nov 2014 22:06:54
 286 B. LEWIS AND G. VON ELBE

mols in the bomb before ignition is given by which is just about to burn and whose com-
bustion will be complete in unit time. S therefore
N u = 11PdRT;. (5)
represents the desired velocity of the flame:
The number of mols left unburned at any
instant is given by N u (1-v;/11). The volume (12)
occupied by this number of mols is (11 -Vb).
By using the gas law once more,
T u , P, and rdrb are obtained from the last three
plots for various values of dr/dt.
Substituting Eq. (5) and transposing, The velocity of the flame in any part of the

v; = 1_ :i (
~: 1_ V~). (7)
vessel, the initial temperature Tu of the gas
about to undergo combustion, the temperature
Tb immediately behind the flame front (from
Eq. (4», and the existing pressure P are now
Since Vb is the volume of a sphere containing the known.
expanded burned gas, its radius rb is given by
ApPLICA TION OF THE METHOD TO GASEOUS
rb -
- _[ 1 -
Pi-Tu(
- 1 -Vi-)]1 , OZONE EXPLOSIONS
ro P Ti 11
We have applied the above system of equations
where ro is the radius of the bomb. to explosions of mixtures of gaseous ozone and
The flame speed is derived from these relation- oXYF:en.
ships as follmvs: From Eq. (2) find ri, the radius Ozone decomposes according to the reaction
of the burned phase when no expansion whatso-
Oa (gas) = It02+t.H,
ever occurs, that is, ri/rO= (v;/V)l. Plot ri against
time, the latter being known from the time- where t.II = 33,920 calories at 18°C and constant
pressure record. Draw three more graphs in pressure. If a spark is passed in gaseous ozone
order to correlate T u , P, and ratio rdrb' plotting or a mixture of ozone and inert gas, an explosion
these variables against rio From the first plot takes place whose intensity is dependent upon
the differential quotient dri/dt is obtained as a the degree of dilution with inert gas. In order to
function of rio As indicated above, dri/dt repre- obtain flame speeds of the order of 102 to lOa cm
sents the thickness of a shell about to be tra- per second the gaseous ozone was diluted with
versed by the flame front in unit time under the pure oxygen, the ratio oxygen/ozone being varied
condition that no expansion of the adjacent from about 1 to 4.
burned phase occurred; that is, the temperature Designating the ratio mols 02/mols Oa by m
and pressure of the shell are Ti and Pi. But by and the energy liberated at constant pressure
expansion the radius of the burned phase has per mol of ozone by t.II, Eq. (4) becomes
increased from ri to rb, the pressure to P, and
the temperature of the unburned phase to T lL •
Tb = (t.II/(1.5+m)CPTu Tb ) + T.. (4a)
The volume of this shell of thickness, dr;/dt and Eq. (1) becomes
before expansion of the burned phase, was
[(1.5+m)/(1+m)]C b T b -Cu T u =K. (1a)
(9)
We shall now present calculations of the
while after expansion the volume of the shell is constant K, using, on the one hand, Eq. (1 a)
dri TuP; with band spectra heat capacities and different
--471"(ri)2- -. (10) combinations of Tb and Tu and, on the other
dt T; P hand, Eq. (2) with experimental pressure and
This volume also equals temperature data from our explosions. For all
practical purposes we may idehtify the mean
(11)
molar heat capacities Cb and CIt with the true
where S IS the actual thickness of the shell molar heat capacities of the burned and unburned

This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
128.248.155.225 On: Sun, 23 Nov 2014 22:06:54
 DETERMINATION OF THE SPEED OF FLAMES 287

gases at the mean temperatures between Tbi TABLE 1. Calculation of flame speed, S, in ozone explosions
and Tbe and Ti and Tue, respectively. Accurate from time-pressure records.
values of the true heat capacity of oxygen are
Time, P, P, S,cm
known. s The heat capacity of ozone is known sec. mm mm ri Tn, T b, per
from band spectra with sufficient accuracy for X 103 Hg Hg OK OK sec.
this purpose. (These were kindly furnished us Experiment 3:13
by Dr. L. S. Kassel.) m=3.054, P i =624 mm Hg, P,=3572 mm Hg, Ti=300.0oK
We shall employ the results of a typical o 624 624 0.599 1.02 300 1239 55.0
explosion, experiment 32, series 3 (3 :32). 20.70 665 630 .601 2.99 301 1240 60.7
27.63 712 650 .603 4.96 303 1241 61.5
Ti=302°K, P,=4814 mm Hg, 34.52 795 700 .615 7.04 309 1247 63.6
37.98 862 785 .637 8.86 317 1253 65.5
Tue=494°K, Pi =492 mm Hg, 41.42 991 885 .666 10.15 327 1259 74.3
T bi =1919°K, m=1.016, 44.88 1141 1070 .702 11.33 343 1274 86.1
48.35 1353 1324 .745 12.41 361 1289 100.0
Tbe = 2069°K, 51.79 1661 1622 .791 13.20 380 1304 114.0
55.25 2155 2024 .845 13.83 401 1322 134.2
m 1 56.97 2499 2560 .907 14.38 427' 1343 158.0
CUTi Tu =--C(02)r/'u+--C(Oa)ri Tu . (13) 58.72 2948 3220 .970 14.80 451 1362 178.0
1+m 1+m 59.41 3192
59.75 3324
60.44 3572
The heat capacities are at constant pressure.
"Yu = 1.328, CPTu/'b, =8.18, CPTi Tbi = 7.935, Tbe = 1373,
c"i:; = (1.016/2.016) X 7.187 + (1/2,016) X 10.50 Tbi= 1239, T u,=462, ri/rb(for P=P i ) =0.599.
=8.83 ca!./deg. mol,
Cb;~~~=9,207 ca!./deg. mo!., Experiment 2 :09
m= 1.497,Pi =367mm Hg,P,=3029 mm Hg, Ti=301.1°K
From Eq. (la)
o 367 370 0.535 1.83 301 1680 141
K = (2.516/2.016) X9.207 X2,069-8.83 X494 8.66 445 389 .546 4.96 305 1684 165
9.70 507 400 .555 6.54 307 1685 187
=19,410 calories, 10.38 564 445 .568 8.20 314 1689 203
K = (2.516/2.016) X9.207 X 1,919 -8.83 X302 11.08 626 575 .610 10.54 333 1704 224
11.77 698 690 .638 11.48 347 1714 237
= 19,370 calories. 12.98 910 820 .672 12.25 361 1726 257
13.57 1045 980 .708 12.90 376 1738 282
From Eq. (2), 14.34 1313 1185 .750 13.46 392 1750 314
14.89 1582 1420 .791 13.87 408 1763 352
RTi[(Pe/P i ) -1J 15.31 1845 1695 .833 14.22 425 1777 395
1= - - - - - - - - - - - 15.62 2110 2000 .878 14.48 441 1790 435
16.17 2627 2825 .978 14.92 477 1818 563
+
RT;( 'Yb - 'Y u) / ('Y u-1) ('Yb -1)K 16.30 2812
16.52 3029
=5266/( -33+0.275K),
"Yu=1.291, CPTu,Tbc=8.445, CPTi Tbi =8.21, T bc =1825,
K = 19,280 calories.
Tbi = 1680, T u, =484, r;jrh(for P =Pi ) =0.531.
The good agreement of the K values shows Experiment 3:32
that Eq. (1) is fulfilled for all combinations of m= 1.016, P i =492 mm Hg, P c=4814 mm Hg, T i =301SK
Tit and Tb and that Eq. (2) meets the require- o 492 495 0.502 0.92 302 1922 160
ments of initial conditions. Therefore values for 5.61 663 510 .508 4.11 304 1923 219
6.30 777 575 .526 7.28 312 1928 287
Cit and Cb based on our explosion data may be 7.00 958 730 .560 10.05 317 1930 309
evaluated by a simple mathematical transfor- 7.52 1164 970 .607 11.76 350 1958 379
7.79 1314 1355 .675 12.95 376 1976 495
mation of the above equations. This procedure 7.97 1469 2042 .765 13.92 411 2003 615
provides an independent method for determining 8.28 1692 2735 .840 14.32 438 2021 693
8.66 2199 3200 .880 14.55 453 2035 733
the heat capacity of ozone.n In all the following 8.94 2700 3760 .925 14.76 468 2044 747
calculations these values of Cu and Cb are 9.11 3134 4440 .975 14.92 486 2063 780
9.32 3702
employed; all values are thus based on data 9.49 4252
obtained from the experiments under consider- 9.59 4542
9.67 4814
ation.
"Yu = 1.276, CPTu,Tbe =8.565, CPTi Tbi = 8.34, Tbe = 2069,
Data on the heat capacity of ozone are presented in
11 Tbi = 1919, T u,=494, r;/rb(for P=P i ) =0.500.
the second paper following, p. 294.

This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
128.248.155.225 On: Sun, 23 Nov 2014 22:06:54
 288 B. LEWIS AND G. VON ELBE

The flame speeds for a considerahle number of where Te is the temperature calculated from the
explosions were calculated, but only tlie results gas law and the maximum pressure P e (see
of three representative ones will be presented, below). After expansion the temperature de-
namely, a moderately slow, a medium, and a creases to Tbi and the pressure to Pi. Then
moderately fast explosion. The results are tabu- the new volume equals
lated in Table I. The radius of the bomb is
150 mm. (15)
The first column for each experiment gives the Combining,
time in milliseconds at which the absolute
pressure in the second column was reached.
and
These values were obtained from the experi-
mental time-pressure record of the explosion.
The next six columns show the results calculated
from the first two columns by the method given
above. The sixth column of the group contains TEMPERATURE GRADIENT IN THE BOMB AND ITs
S, the flame speed relative to the moving gases EFFECT ON TfIE DETERMINATION OF
SPECIFIC HEATS BY THE Ex-
at various radii, rb, given in column 3 of this
group. In column 5 of this group the corre- PLOSION METHOD

sponding T b , calculated from Eq. (4a), is given Before developing the above treatment we
for convenience. The CPT"Tb in this equation attempted to devise a simplified method for
may be found readily by interpolating between determining flame speeds by assuming that the
the two limits given for CPTue Tbe and CPTi 7'bi, temperature is equalized throughout the burned
which are indicated under each experiment in phase at any time. However, the values of Tb
Table I and which were obtained from earlier calculated and the values of Tb found from
work. 5 Eq. (4a) varied widely, indicating a rather steep
ri/rb at the beginning cannot be determined temperature gradient from the center of the
easily from the above equations because both bomb to the wall which could not be neglected.
values approach zero. It is evaluated in the Mache 9 developed a formula for the tempera-
following way. Assume that a very small number ture gradient in a spherical bomb at the instant
of mols, dn, burns at the beginning. The volume
combustion is complete. If Te(vi/V) is the temper-
it occupies, if no expansion occurs, equals
ature at the surface of the sphere which encloses
Vi = dn(RTe/ P e), (14) a fraction, Vi/ V, of the total gas burned, then

Te(vt/v)=--
'Yb-l l+m (Pe)('Yb-l)hl --+--
Ti -
K 'Yu (P)('Yu-1)h,,]
- , (18)
'Yb 1.5+m P RTi '1',,-1 Pi

in which all the symbols are known. pressure recorded is a little less than the maxi-
In the explosion method for determining the mum pressure which would have been recorded
heat capacities of gases the maximum pressure had the temperature throughout the bomb been
developed is used to calculate the temperature equalized without heat loss. We have made
of the gases, T e, from the gas law under the extended calculations of this effect for the
assumption that the temperature is equalized present type of explosion including the effect of
throughout the bomb. As already pointed out by dissociation and find that the difference between
Hopkinson 1o this procedure would be strictly these pressures is from less than 0.1 percent to
justified if the heat capacity of the gases did less than 0.3 percent in the extreme case of the
not vary with temperature. In the normal case, highest temperature explosions. This work will
however, where one deals with polyatomic be published soon.
molecules and their dissociations the maximum For other types of explosive mixtures one may

This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
128.248.155.225 On: Sun, 23 Nov 2014 22:06:54
 DETERMINATION OF THE SPEED OF FLAMES 289

estimate the effect to be very small except in

explosions where large dissociation occurs. But
even in such cases the pressure difference can be
calculated and therefore this effect does not
constitute an objection to the explosion method.
Fig. 1 shows graphically the conditions in the
bomb while combustion is in progress and before
and after combustion for explosion 2 :09. The
hemisphere represents one-half of the bomb. The
abscissa is the radius of the bomb, 15.00 cm.
The upper quadrant shows the positions reached
by the flame front after equal time intervals.
The time interval in this instance is 1.65 X 10-3
sec., which is exactly 1/10 of the total explosion
time. The pressure existing in the bomb for
these positions of the flame front are indicated
on the left. Temperatures 1'" and Tb in front of
and behind the flame front, respectively, are
given to the right of these. Thus the upp'er
quadrant provides a complete picture of condi-
tions during the progress of the flame. It should
be emphasized that temperature Tu in front of
the flame is equalized throughout the entire
unburned phase, whereas temperature Tb behind
the flame is localized in the immediate vicinity
of the flame front. From the rear of each flame-
front position the temperature rises toward the
center of the bomb. The arrows represent flame
1825"!----
speeds S relative to the moving gas when the
flame reaches the positions indicated. FIG. 1. Experiment 2 :09. Showing the hemisphere of the
The solid curves in the lower quadrant desig- bomb. Abscissa = radius of bomb in cm. Total explosion
time 16.SXlO-3 sec. Upper quadrant: Contains speed of
nated by numerals with the subscript i, which flame in various parts of bomb. Curves marked tb, 2b, etc.,
are volumes occupied by gas burned in equal time intervals,
denotes pre-ignition conditions, mark the bound- in this case l/tO of total explosion time. Pressure in bomb,
aries of the spheres occupied by the same mass temperature Tu in unburned phase and temperature Tb
of gas contained in the corresponding spheres in immediately behind flame front when flame reaches points
indicated after successive tenths of total explosion time, are·
the upper quadrant whose boundaries are marked given to the left. Lower quadrant: Solid curves, ti, etc., are
by numerals with the subscript b. The difference volumes occupied by same mass of gas enclosed in lb, etc.,
for condition that no expansion of the gases occurred.
between the corresponding i and b spheres Dotted curves le, etc., are final volumes occupied by same
results from the expansion which occurs on mass of gas in li, etc., or lb, etc., after lb, etc., has been
compressed due to subsequent burning of the outer shells.
combustion. The two-shaded portions represent Temperature gradient existing in bomb when combustion
this for the mass of gas burned during the first is complete, is given to the left.
1/10 of the total explosion time.
The dotted curves designated by numerals meters of the dotted spheres were calculated from
with the subscript e mark the boundaries of the the formula given by Mache, which by a simple
spheres occupied by the same masses of gases transformation may conveniently be written
corresponding to the i and b states after the
(19)
explosion is complete or when the maximum
pressure is reached. They show the extent to The temperatures existing at the boundaries of
which the b spheres have been compressed by these spheres after combustion is complete may
the subsequent combustion. The radii in centi- be calculated from Eq. (18) for the same values

This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
128.248.155.225 On: Sun, 23 Nov 2014 22:06:54
 290 B. LEWIS AND G. VON ELBE

of P used in Eq. (19). These temperatures,

given to the left of the lower quadrant, may be
compared with temperature Te = 2072° K, which
would exist if the temperature were equalized.
The temperatures of the inner portions of gas
are somewhat too high because dissociation of
oxygen has been neglected in these calculations,
and therefore the value of Cb used is somewhat I I I I I I
o 3.3 6.6 9.9 13.2 16.5 19.8 22.1 25.4xlO-3 sec.
too small. Radii reo however, would be affected TIME~
only slightly. We believe the picture, as pre- FIG. 2. Time-pressure record, explosion 2 :09.
sented, is nevertheless substantially correct.
Table II contains the numerical data shown
in Fig. 1. I t is of interest that in the first part of the
Fig. 2 is a photograph of the time-pressure record the curve practically coincides with the
record of explosion 2 :09 from which Fig. 1 was zero line. It has been stated in the literature
derived. that this circumstance is due to the existence of a
time lag between the passage of the spark and
TABLE II. Numerical data for experiment 2 :09 used in Fig. 1.
ignition. However, this conclusion is not ad-
Time, S,em mjssible. 12 In the present case, for example, ref-
sec. rio 'i 'b, T u• Th, P, per T e• r"
X 10 3 em em oK oK mmHg seC'. OK em
rb
erence to Figs. 1 and 2 shows that when 1/5 of
0 0 0.531 0 301 1680 367 (140) 2670 0
1.65 .79 .534 1.48 302 1680 369 146 2667 .86 the total explosion t.ime has elapsed the pressure
3.30 1.62 .538 3.01 303 1681 374 151 2657 1.75
4.95 2.52 .544 4.63 305 1683 385 163 2644 2.71 rise becomes just readable at 7 mm Hg. In this
6.60 3.52 .554 6.36 307 1687 400 182 2633 3.76
8.25 4.65 .569 8.17 314 1692 447 202 2571 4.97 same time the flame front has progressed 3
9.90 5.91 .595 9.94 329 1701 530 218 2496 6.32
11.55 7.36 .640 11.50 347 1716 685 237 2380 7.81 centimeters or 1/5 of the radius of the bomb.
13.20 9.10 .708 12.85 376 1738 980 281 2225 9.58
14.85 11.45 .815 14.05 416 1776 1560 370 2013 11.76
16.50 15.00 1.000 15.00 484 1825 3029 (650) 1825 15.00
12 Payman (Fuel 1, 36 (1922)) also realized this.

This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
128.248.155.225 On: Sun, 23 Nov 2014 22:06:54