One-Dimensional Organic Conductors TTF-TCNQ and other Organic
Semi-metals," Energy and Charge Transfer in Organic Semiconductors,
(Eds. K. Masuda and M. Silver), Plenum Press (1974).
39-f. O. Poehler, A. N. Bloch, T. F. Corruthers, and D. O. Cowan,
"The Organic Metallic State: Some Physical Aspects and Chemical
Trends," Proc. NATO Conference on Chemistry and Physics of OneDimensional Metals (Ed. H. J. Keller), Plenum Press (1977).
4OR. S. Potember, T. O. Poehler, and D. O. Cowan,"Electrical Switch-ing and Memory Phenomena in Cu-TCNQ Thin Films," Appl. Phys.
Lett. 34, p. 405 (1979).
and reaction intermediates, to the detection of free
radicals stabilized at very low temperatures, and to the
ionization of substances by electron impact.
He has served as a member of the NAS/NRC Advisory Committee for the Army Research Office and an Advisor to NATO's Scientific Affairs Division. In 1954,
Dr. Foner received the Physical Sciences Award of the
Washington Academy of Sciences for work in free
radical chemistry and physics. He is a member of the
Combustion Institute, the Philosophical Society of
Washington, and a Fellow of the AAAS, the American
Physical Society, and the Washington Academy of
Sciences.
Robert W. Hart (left) and Samuel N. Foner
SAMUEL N. FONER is Vice-Chairman of the Milton
S. Eisenhower Research Center and Supervisor of its
Electronic Physics Group. Born in New York City
(1920), he studied physics and mathematics at what is
now the Carnegie-Mellon University, where he received
his D.Sc. degree in physics in 1945. He was employed as
an instructor in the Physics Department and later as a
research associate of the Manhattan Project, working in
the laboratory of the Nobelist Otto Stern who instilled in
him the use of conceptually simple experiments to
answer complex questions.
Dr. Foner joined APL in 1945 and became associated
with the Research Center as Supervisor of the Mass
Spectrometry Group (1947-52) and the Electronic Physics
Group (1953-present). He has made many noteworthy
contributions to the mass spectrometry of free radicals
ROBERT W. HART is Chairman of the Milton S.
Eisenhower Research Center and Assistant Director of
APL for Exploratory Development. Born in Yankton,
SD, in 1922, he studied at the University of Iowa and
received his Ph.D. degree in physics from the University
of Pittsburgh in 1949. After a year of teaching at the
Catholic University in Washington, he joined APL in
1950. He has been a member of the Research Center
ever since.
During the 1960s, Dr. Hart developed a detailed
theory of the complex combustion behavior of solid propellants in rockets in collaboration with the late Frank
T. McClure and as member of the Joint Armed Services
Committee on Combustion Instability. As Supervisor of
the Special Problems Research Group (1954-1975), his
interests covered theoretical aspects of wave scattering,
the structure of the eye, and other phYSical and
biophysical topics.
Dr. Hart is a member of the American Physical Society and of the Combustion Institute. Outside of professional activities, he is interested in the origin and evolution of civilization and science.
RESEARCH RETROSPECTIVES
THE STRUCTURE OF FLAMES
Flames have been the most important source of
heat, light, and power since the earliest days of
civilization. At present, the combustion of fuels is,
by far, the largest chemical operation under human
control. Yet, until quite recently, detailed knowledge of what goes on within a flame did not exist.
Although the complexity of combustion is not entirely understood, even today, what was virtually
terra incognita has been opened up during the past
25 years by the classic studies at APL by Robert
M. Fristrom, Arthur A. Westenberg, and their colleagues.
What does one need to know about a flame?
Chemists want a detailed accounting of the steps by
which fuels (such as oil, natural gas, or coal) and
oxidizers (such as air) are converted into products
of combustion (water, oxides of carbon, soots, and
January- March 1980
ash) as well as the intermediate reaction paths that
are involved in this transformation. They want to
understand why and how inhibitors can extinguish
flames or prevent engine "knock" and know how
rapidly these transformations can take place.
Physicists, on the other hand, are interested in
temperature effects, radiation, the flow fields set
up by the gases moving into and out of the flames,
and countless other physical properties.
What sets flames apart from more conventional
chemical transformations is that one is dealing with
a very intricate situation in which chemical reactions are closely coupled with the physical flow of
substances into and out of a reaction zone, accompanied by a steep rise in temperature, abrupt
changes in composition, and numerous optical and
electrical phenomena that may be important under
specific circumstances. In a distance of less than 1
mm, temperatures can change by thousands of
33
degrees and gas concentrations can rise or fall
abruptly. The thinness of this transformation zone
held back experimenters in exploring the detailed
structure of flames.
A fundamental advance in understanding the
simplest combustion case (where gaseous reactants
are mixed prior to combustion and no solid reaction products are formed), came about in the early
1950's. First, a theoretical analysis of flames was
made l whereby simplified models of combustion
processes could be analyzed in detail. This led to
suggestions of how one could, in principle, separate
the chemical transformations from the simultaneous physical processes (mainly diffusional). Second, it was found that most flames, when stabilized at pressures well below atmospheric, would
widen in thickness without altering the sequence of
the chemical transformations. This made it possible
to introduce sampling probes and thermocouples
into the reaction zone and obtain point-by-point
samples for subsequent analysis of the concentration of reactants, intermediates, and products as
well as temperature profiles.
APL's work pioneered in the development of
these experimental tools and their application to
the analysis of simple flame systems. Mass spectrometers were found to be useful in identifying the
chemical species that survived the sampling process.
For highly reactive intermediates such as radicals
and atoms, stablizing techniques were developed to
preserve them for later analysis. Tiny microprobes
and thermocouples were built to obtain spatial
resolution in flames whose thickness now extended
over several millimeters. 2
Once the value and workability of these techniques were recognized, they were quickly adopted
by others. A vigorous worldwide exploration of
many flame systems was started. Much was learned
about the intricate interplay by which stable reactants transform into stable products by way of intermediate steps and, in particular, about the crucial role played by free radicals in mediating the
rapid transformations.
The picture that has emerged 3 of the structure
of, for example, a hydrogen-oxygen flame with excess hydrogen is that initially, as the reactant gases
enter the combustion region, they are heated somewhat by conduction as heat flows toward the incoming gases from the hot combustion products. A
little later, heat-producing chemical reactions commence. They are initiated by reactive free radicals
that diffuse from the hot side of the flame, where
they are generated, toward the incoming gases and
attack hydrogen molecules, leading to the formation (as well as the disappearance) of the hydroperoxo radical (H0 2 ) and water. At a still later
(and hotter) stage, the relatively slow reaction 0 +
H2~ OH + H becomes important. Because more
free radicals are generated than are consumed, this
reaction stage is the source of the reactive free
radicals that are active earlier in the flame. Finally,
34
after the main rapid reactions have taken place, a
sorting-out zone follows where the free radicals
that were formed in excess amounts in the previous
stage recombine relatively slowly until the system
settles down to an equilibrium state in which reactions respond only to the slow temperature changes
brought about by heat losses.
For the "simple" hydrogen-oxygen flame, where
most of the possible intermediate steps have now
been identified and their reaction rates measured,
the overall behavior of the flame can be described
by a quite limited number of reaction steps. 4 For
hydrocarbon flames, many more chemical interactions are possible because of the presence of carbon
in the fuel molecule. An overall description of hydrocarbon flame behavior based on individual steps
is now nearly in hand. Here, too, the role of free
radicals as crucial mediators (hydroxyl radicals if
the flames are deficient in fuel for complete combustion, or hydrogen atoms if there is an excess of
fuel) is beyond doubt.
During the combustion of hydrocarbons with excess air, the oxidation of hydrogen and carbon
monoxide (which appear as intermediates in the
flame) is of particular importance. The hydrogen
oxidation furnishes the oxygen and hydrogen atoms
and hydroxyl radicals that attack the hydrocarbon
molecule to form methyl (CH 3 ) radicals. The latter
subsequently interact with oxygen to form formaldehyde, which reacts in further steps to produce
carbon monoxide and, ultimately, carbon dioxide.
This sequential formation and disappearance of intermediates (carbon monoxide, hydrogen, formaldehyde, and atoms and radicals) was clearly shown
in the experiments of Fristrom and Westenberg S
and is fully supported by the detailed reaction
scheme proposed for hydrocarbon oxidation in
flames 6 (Fig. 1).
In fuel-rich flames, many more intermediate
steps are possible. Polymerization reactions that
lead to the temporary appearance of hydrocarbons
that are of higher molecular weight than the initial
fuel become important. They give rise, subsequently, to aldehydes and unsaturated hydrocarbons. A
complete description of all the individual reaction
steps is not yet possible because of the lack of reaction rate data for all of the numerous elementary
reactions involved.
One goal of the flame structure work was to obtain quantitative information about individual reaction steps. This turned out to be difficult in practice because so many reactions are proceeding simultaneously even in the simplest flame that it is
nearly impossible to single out anyone for detailed
analysis-with one important exception. In the hot
stream of combustion products beyond the active
reaction zone, free radicals are still present in
measurable amounts. This region can be used as a
"hot bath" where the reactions of compounds with
free radicals can be studied. It became evident that
it was possible to inject traces of well-known flame
Johns HopkinsAPL Technical Digest
0.90
c:
o
.~
0
0.15
H2O
0. 85
c:
<I>
(,,)
c:
c:
0
o(,,)
I
.~
v 0.10
c:
u
<I>
0.80
'0
:ii
6
IN
*-'0
..s
cO 2
0.05
6
°2
u
0'
u
g
8
0'
0.75
CO
CH 4
N
I
'0
~ 0.004
1jN 0.0021:1....--"---:
2000 r r - - --
r - - - - - - - , - - --
H2
CH 20
--r-----,------,
niques at lower temperature. 7 ,8 For the first time it
was demonstrated that these measurements give
concordant results.
Many intriguing and important problems remain,
especially in applied areas where fuel! oxidizer mixing limitations, catalytic surface effects, soot formation, and many other subtle interactions may
lead to undesirable end effects. However, for the
central problem of gas phase combustion, 200 years
of flame research have, at long last, brought about
a remarkable confluence of theory and experiment.
Taking into account the dominant physical processes of diffusion and heat conduction and the
numerous interacting chemical reaction steps, the
structure of flames can now be viewed in its full intricacy. Complex flame systems can be constructed
out of the many individual reactions that proceed
within a flame and the overall behavior of such
flames can, in principle, be predicted.
Q
-; 1500
WALTERG.BERL
~
~
REFERENCES
1000
~
~
u.
500
Distance along f lame (mm)
Fig. I-In 1960, a low pressure fuel-lean flame of methane and
air was analyzed by Fristrom et als for the appearance and
disappearance of chemical species and for local temperature.
The reaction zone was approximately 4 mm thick. Experimental
results are shown as individual data points. Recently, J . Warnatz 6 was able to calculate composition and temperature profiles of the same flame from known reaction rates of approximately 80 of the important individual steps that are postulated
to occur during the oxidation of methane (solid lines) . The
agreement between prediction and experiment is excellent.
inhibitors (methyl halides) into this bath, determine
their rates of reaction with hydrogen atoms at the
high temperature, and compare the results with extrapolations obtained with entirely different tech-
THE SEARCH FOR H0 2
In more than two hundred years, chemists have
isolated and identified about 100 chemical elements
and millions of compounds into which these
elementary building blocks can be combined. This
continuing and unending quest (the number of
potential combinations of elements into compounds
is virtually limitless) has been accompanied by an
intensive effort to learn more about the bonds that
hold these building blocks together in recognizable
structures and shapes and to discover the rules that
determine the rates and the pathways by which one
chemical structure changes into another.
Until about 50 years ago, the study of chemistry
was based on a belief in stability. To be sure, many
January-March 1980
IJ. O. H irschfelder, C. F. Curtiss, and D. E. Campbell, "The Theory of
Flame Pro paga tion," J . Phys. Chern. 57, pp. 403-4 14 (1953) .
2R . M. Fris trom and A . A. Westenberg, Flame Structure, McG raw- H ill
(1965) .
3G. Dixon Lewis, "Kinetic Mechanis ms, Structu re, a nd Properties of
Premixed Flames in H ydrogen-Oxygen- Nit rogen Mixtures," Phil. Trans.
Roy. Soc. 292 , pp. 45-99 ( 1979).
4 J . Warnatz, "Calculation of the Structure of Laminar Flat Flames II:
Fla me Velocity and Structure of Freely P ropagating Hydrogen-Oxygen
and Hydrogen-Air Fla mes," Ber. Bunsenges. Phys. Chern. 82, pp .
643 -649 ( 1978) .
SR. M . Fristrom, C. Gru nfelder, and S. Favi n, "Methane-Oxygen Flame
Struct ure I: C haracteristic P rofiles in the Low- P ressure, Laminar, Lean,
Premixed Methane-Oxygen Flame," 1. Phys. Chern. 64 pp.1386-1392
(1960)
6J . Warnatz, "Flame Velocity and Structure of Laminar H ydrocar bon
Flames," Proc., Seventh International Col/oq. on Gas Dynamics of Explosions and Reactive Systems (to be pu blished) .
7L. W. Ha rt , C. G run fe lder , a nd R. M. Frislrom, "The ' P oint So urce'
Techniq ue Using Upstream Sampling fo r Rate Constant Determi nat ions
in Flame Gases," Combust. Flame 23 , pp. 109-119 (1974)
8A. A. Westenberg and N. de Haas, " Rates of H + C H 3X Reactions,"
J. Chem. Phys. 62 , pp.3321-3325 (1975).
levels of stability were identified and techniques
were developed to move from one level to another.
Some substances were so labile that they would
barely survive at room temperature. At high
temperatures most compounds would change into a
relatively small number of stable products. A few
elements like radium and polonium showed signs of
instability. But once substances such as hydrogen
(H 2) and oxygen (0 2) molecules reacted with each
other, they were expected to form only water or,
on occasion, hydrogen peroxide (H 20 2). The
details of how such transformations take place
were but dimly perceived. It was generally assumed
that a direct reaction (commonly written as 2H2 +
O 2 -. 2H 20) occurs that involves no other
chemical species.
35
A change in this simplistic view came about in
the 1920's. Researchers working in the area between physics and chemistry became familiar with
the details of explosions and flames where chemical
reactions proceed at speeds well beyond the leisurely pace of the conventional chemical reaction.
Puzzling observations were made. For example, if a
container is filled with hydrogen and oxygen at a
high enough temperature, its contents would always
explode during mixing. But by lowering the temperature below a critical level, the reaction between
the two gases settles down to a slow rate of water
formation that might take hours to go to completion. However, when one starts to withdraw such a
slowly reacting mixture from the container by
means of a vacuum pump, the remaining mixture
of hydrogen, oxygen, and water vapor suddenly
(and always at a precisely reproducible pressure
designated the "explosion limit") explodes with a
bright flash and the reaction is completed instantly.
This extraordinary behavior of a sudden transition from a slow reaction to a very fast one by a
mere change in pressure became a turning point in
the interpretation of chemical reactions. A general
explanation was provided by the Russian chemist,
N. N. Semenov, 1 who proposed that highly reactive
atoms and so-called "free" radicals mediate the
transformation among the reactants. These intermediaries, whose concentration depended on the
particular conditions of the experiment, were, in
fact, crucial participants in the reaction. Their very
reactivity made it impossible (at the then-existing
state of the art) to detect or to isolate them by conventional chemical means. Because of their inaccessibility, they had not been thought of by chemists as being important in chemical reactions even
though physicists had been aware of their existence
in many experiments in which gases were exposed
to either high temperatures or electric discharges.
For the hydrogen-oxygen system, not many
choices of intermediaries are possible. Hydrogen
atoms (H), oxygen atoms (0), and hydroxyl
radicals (OH) were already known in physics. One
could postulate that water is formed by the elementary reaction
OH + H2 -+ H 20 + H,
(1)
in which one reactive substance (OH) disappears
but another one (H) appears. This atom, in turn,
would react with oxygen
H + O 2 ~ OH + O.
(2)
The OH formed in (2) would react as in (1) with
more hydrogen to form another molecule of water.
The oxygen atom, on the other hand, can react
with a hydrogen molecule in the following way:
o + H2 ~ OH + H.
(3)
According to this scheme, once a reactive
substance is provided to initiate the process, water
can be formed in a simple series of chain reactions
in which the key radicals are regenerated endlessly.
In fact, in reactions (2) and (3) the number of
"chain carriers" increases in each reaction step.
36
However, these three reactions are inadequate by
themselves to explain the observed events. Since
more chain carriers are formed than are used up,
the formation of water would be expected to
become progressively faster with time and always
end up in an explosion. A way had to be proposed
to remove the chain carriers as quickly as they are
formed in the region where the reaction proceeds
slowly and to let them build up at a rapidly increasing rate only at the explosion limit.
A careful analysis of the experimental observations on how the explosion limit is affected by gas
temperature, container size, pressure, mixture ratio,
and the presence of inert diluents led to a reaction
scheme that forced the postulation of one additional intermediary, the hydroperoxo radical
(H0 2), which has the extraordinary properties of
being a stable enough free radical to compete with
the chain reaction involving hydrogen atoms (reaction (2» but also unstable enough to be destroyed
once it reaches the wall of the reaction vessel. Only
by postulating its formation by the reaction
H+0 2 + M-+ H0 2 + M
(4)
(where M can be any of the stable participants in
the reaction) was it possible to account for the
observed system behavior in detail. One of the proponents of H0 2 said about its existence "These
conclusions are inescapable.,, 2
When this "inescapable" interpretation was proposed in the 1930's, there was no experimental
evidence from any source for the physical existence
of H0 2 nor would there be for 20 years. Yet after
studying reactions involving the oxidation of hydrogen-containing materials that included the entire
family of hydrocarbon fuels and of myriad other
organic compounds, physical chemists had few
doubts that the proposed free radical must exist.
Without H0 2, the observed system behavior simply
could not be interpreted; with it, the observations
became understandable.
In the middle 1940's there was a brief report
from the Shell Research Laboratories that when a
hydrogen/ oxygen flame is placed in front of the
sampling port of a mass spectrometer, it would
produce a molecular fragment of the correct mass
(33) attributable to H0 2. But the evidence was
flimsy. The system was too complex and too difficult to analyze. Convincing experimental evidence
for H0 2 was still lacking.
It was another 10 years before the definitive experiments that would identify H0 2 were made by
Samuel N. Foner and Richard L. Hudson at APL
in 1953. 3 This was done by combining a sensitive
mass spectrometer with a reliable "molecular
beam" inlet and a reacting system that could leave
little doubt about the events that were occurring.
The reaction scheme was to generate hydrogen
atoms in a separate electric discharge, mix them
rapidly with oxygen molecules before all the
hydrogen atoms had time to recombine, and
analyze the reaction products for H0 2. In this inJohns Hopkins A PL Technical Digest
itial effort, the mass spectrometer, despite its sensitivity, was hard pressed to show a peak at mass
33 caused by H0 2 • The peak was there, nevertheless, disappearing when the hydrogen atoms
were shut off and reappearing when they were
turned on again (Fig. 1).
Once such a trailbreaking discovery is made, a
veritable flood of additional information commonly comes to light quickly. 4 Other reactions were
found (mainly involving hydrogen peroxide) that
proved to be more convenient sources of H0 2 in
much higher concentration. Many rates of H0 2
reactions with itself, with other simple molecules,
and with hydrocarbons have now been measured. 5
The spectra of H0 2 were obtained, from which
many of its structural properties have been
estimated. 6 Its terrestrial existence is beyond question. Only its presence in interstellar gas clouds,
where so many other free radicals have recently
been discovered, has yet to be established. 7
Nearly 20 years later, H0 2 made one more appearance in the APL research effort. The reaction
of H0 2 with carbon monoxide was proposed at
one time as an important link in atmospheric
chemistry research. If the reaction proceeded rapidly, it would offer an attractive pathway for
eliminating the carbon monoxide that is generated
when fossil fuels are burned and for whose
scavenging from the atmosphere no really satisfac-
Discharge on
tory mechanism had been known. In a series of
elegant experiments, 8 A. Westen berg and Newman
de Haas measured free radical reactions at low
pressure in simple systems in which the rates of appearance or disappearance of these radicals could
be determined with great precision by using electron spin resonance techniques. The hydrogen
atom/ oxygen molecule reaction was a seemingly
straightforward pathway that, in the presence of
carbon monoxide, should permit rate measurements
of the latter with H0 2 • The results indicated a very
fast reaction, quite in contrast to earlier, more indirect results deduced from more conventional explosion experiments.
In order to resolve this conflict, people elsewhere
(particularly at the University of Maryland) carried
out independent measurements or the H0 2 -CO
reaction using isotope tracer techniques and higher
pressures. 9 There is now a consensus that H0 2 is
not as effective a scavenger of carbon monoxide as
was believed by Westenberg, even though its role in
reacting quickly with other atmospheric contaminants such as nitric oxide and sulfur dioxide is
important. H0 2 may have formed in Westenberg's
experiment so as to produce a nonequilibrated
species with unusually high energy content and
great reactivity. If this is so, a new chapter of reaction research is opening where reactants are not in
equilibrium with their surroundings. Such nonequilibrium conditions may prevail at high altitudes
(low pressure) or in combustion situations where
reactions occur at very high speed.
The detection of H0 2 validated a theoretical
prediction of great subtlety. For two decades experimenters were challenged to devise experiments
that would provide convincing evidence. Foner and
Hudson met the challenge.
WALTERG. BERL
REFERENCES
T ime (one minute intervals)
Fig. I-The mass spectrometric ion intensities for (a) mass
number 32 (due to 0 16 0 16 ), (b) mass number 33 (due to 0 16 0 17
and HO I6016) , (c) mass number 34 (due to 0 16 0 18 and
HH01 60 16) show that the pronounced plateaus for mass 33
must have been due to a molecular fragment (H02) that is produced only during the period when hydrogen atoms generated in
an electric discharge subsequently react with oxygen. Much of
the intensity background at mass 33 and 34 comes from the
ever-present oxygen isotopes 0 16 0 17 and 0 16 0 18 . Since no
notable changes at mass 34 are observed with the discharge on,
hydrogen peroxide was not formed in appreciable amounts .
Consequently, its fragments could not have been responsible for
any of the peaks at mass 33 .
January-March 1980
IN . N. Semenov, Chemical Kinetics and Chain Reactions, Oxford
(1935).
28. Lewis and G. Von Elbe, Combustion, Flames, and Explosions of
Gases, Cambridge (1938).
3S. N. Foner and R. L. Hudson, "Detection of the H02 Radical by
Mass Spectrometry," J. Chern. Phys. 21, pp. 1608-1609 (1953).
4S. N. Foner and R. L. Hudson, "Mass Spectrometry of the H02 Free
Radical," J. Chern. Phys. 36, pp. 2681-2688 (1962).
5A. C. Lloyd, "Evaluation and Estimated Kinetic Data for Phase Reactions of the Hydroperoxyl Radical," Int. J. Chern. Kin. 6, pp. 169-228
(1974).
6T. T. Pankert and H. J. Johnson, " Spectra and Kinetics of
Hydroperoxy1 Free Radicals in the Gas Phase," J. Chern . Phys. 56, pp.
2824-2838 (1972).
7S. Saito, "Microwave Spectroscopy of Short-Lived Molecules," Proc.
XXVIIth International Congress of Pure and Applied Chemistry: Vol.
2, Physical Chemistry, pp. 1239-1250 (1977).
8A. A. Westenberg and N. de Haas, "Steady State Intermediate Concentration and Rate Constants, Some H02 Results," J . Phys. Chern. 76,
pp. 1586-1593 (1972) .
9D. D. Davis, W. A. Payne, and L. J . Stief, "The Hydroperoxyl
Radical in Atmospheric Chemical Dynamics: Reaction with Carbon
Monoxide," Science 179, pp. 280-282 (1973).
37