General Considerations: Nomenclature

1
General Considerations
Nomenclature
Dimensional Quantities
Description S.I. Units
a Mean molecular velocity. Sound speed m s−1
cp Specific heat at constant pressure J K−1 kg−1
cv Specific heat at constant volume J K−1 kg−1
dL Scale of laminar flame thickness m
D Normal propagation velocity m s−1
D Molecular diffusivity m2 s−1
e Energy J ≡ kg (m/s)2
E Activation energy J mole−1
EG Gamow energy J
kB Boltzmann’s constant J K−1
Molecular mean free path m
L Length of tube, burner or space scale m
m Mass flux kg m−2 s−1
M Mass of the sun kg
n Number density m−3
p Pressure Pa
qm Heat of combustion per unit mass J kg−1 ≡ (m/s)2
R Radius m
s Entropy J K−1 kg−1
S Surface area m2
t Time s
T Temperature K
T∗ Cutoff temperature K
UL Laminar flame speed m s−1
Ub Laminar flame speed w.r.t. burnt gas m s−1
v Impact velocity m s−1
13
Downloaded from https:/www.cambridge.org/core. UCL, Institute of Education, on 15 Jan 2017 at 12:15:15, subject to the Cambridge Core terms of use, available at
https:/www.cambridge.org/core/terms. https://doi.org/10.1017/CBO9781316162453.003
14 General Considerations
ρ Density kg m−3
σo Collision cross section m2
σnr Collision cross section for nuclear reaction m2
τ A characteristic time s
Nondimensional Quantities and Abbreviations

cst. Constant
M Mach number U/a
Z Mixture fraction
γ Ratio of specific heats cp /cv
ϑ Stoichiometric coefficient
φ Equivalence ratio
CJ Chapman–Jouget
ICF Inertial confinement fusion
DDT Deflagration-to-detonation transition
D-T Deuterium-tritium
SNI Supernovae type I
SNII Supernovae type II
ZND Zeldovich–von Neumann–Döring
Superscripts, Subscripts and Math Accents

a∗ Critical value
ab Burnt gas
acoll Collisions
aCJ Chapman–Jouget conditions
adiff Diffusion
ae Elecron
anr Nuclear reaction
aN Neumann state (behind a shock)
ar Reaction
au Unburnt gas
1.1 Introductory Remarks

In this introductory chapter we briefly introduce the physical background and the context
of the phenomena that are studied in this book. The discussion is limited to orders of
magnitude and dimensional analysis.
As first demonstrated by the experiments of Lavoisier at the end of the eighteenth
century, combustion is an exothermic chemical reaction between a fuel, such as hydrogen
or a hydrocarbon, and an oxidant, generally the oxygen of ambient air. The reaction rate is a
strongly increasing function of temperature, leading to thermal self-acceleration. Due to the
1.2 Combustion Waves on Earth 15
thermal sensitivity, combustion occurs almost exclusively in thermal fronts. The concept of
a well-stirred (homogeneous) reactor, widely used by chemists, is of only limited interest
in practical combustion. Self-propagating exothermic reaction waves, flames (subsonic
waves called also deflagrations) and detonations (supersonic waves) were identified in
the nineteenth century. Self-consistent theoretical analyses of the dynamics of such waves
and of the instabilities of their fronts started in the mid-twentieth century and are still in
progress. The kinetics of heat release in combustion is complex; it involves hundreds of
elementary reactions and tens of species. Examples are given in Chapter 5, while the basic
elements of chemistry are recalled in Chapter 14. Hopefully most of the phenomena can be
described using ultra-simplified chemistry schemes.
The dynamics of combustion waves is a key issue in different areas of physics. For
example the deflagration-to-detonation transition, which is not yet fully understood, con-
cerns not only the safety of power plants but also the explosion of stars in astrophysics.
With the development of nuclear physics during the twentieth century, astrophysicists
explained how stars can maintain a quasi-stationary state of high temperature thanks to
the energy released by nuclear fusion reactions, compensating the losses by radiation.
These reactive phenomena have some analogies with ordinary combustion, but have also
differences, arising mainly from the extreme conditions of temperature and density at which
thermonuclear reactions occur. A short introduction to thermonuclear fusion is given in
Section 1.3.1. In contrast to combustion on earth, the dynamics of reactive fronts in such
extreme conditions is not yet a mature field, and a detailed presentation is outside the
scope of this book. The explosion of stars during the gravitational collapse at the end of
their lifetime, the famous supernovae that are typically 109 –1010 times brighter than the
sun, are fascinating phenomena that are not fully explained. Shock formation and/or the
deflagration-to-detonation transition are expected to play essential roles in the outbursts
of supernovae. In this book, which is mainly concerned with combustion on earth, these
phenomena will be discussed in simple physical terms with the objective of extracting the
basic mechanisms without entering into the details of nuclear physics.
Ablation fronts, which arise in inertial confinement fusion (ICF), will be discussed in
the same spirit. The concept of this method of energy production, whose feasibility is not
yet proved, is to burn a few milligrammes of nuclear fuel by imploding a spherical shell
using high-power laser radiation. The ignition of a nuclear reaction at the central hot spot,
compressed to more than 1000 times the density of liquids, depends critically on the control
of hydrodynamic instabilities of the ablation front during the implosion. The similarities
and differences with gaseous flames open new perspectives in the study of the dynamics of
such thermal fronts.
1.2 Combustion Waves on Earth

1.2.1 Combustion Modes
Two different modes of combustion are identified according to whether the fuel and the
oxidant are initially mixed at the molecular level or spatially separated. The former is called
premixed combustion and the latter non-premixed.
Figure 1.1 Planar flame propagating from the open to the closed end of a tube. The unburnt gas is at
rest, and the flame propagates at a speed UL with respect to the tube. The hot burnt gas flows out of
the tube with velocity UL (Tb − Tu )/Tu ; see (2.1.2).
Propagating Waves
In premixed combustion (gaseous, liquid or solid), the reaction generally propagates
through the initial mixture as a reaction wave, transforming reactants into combustion
products. There are two very distinct types of propagating combustion waves, depending
on the conditions of ignition:
• Premixed flames are very subsonic and quasi-isobaric since their propagation is governed
by thermal conduction, and the term deflagration referred usually to fast flames essen-
tially in turbulent flows.
• Detonations are supersonic waves composed of a shock wave with a strong pressure
pulse, followed by a reaction zone.
These waves are thin, of a few tenths of a millimetre thick for flames and a few
millimetres to centimetres for gaseous detonations in ordinary conditions. They are thus
often assimilated to infinitely thin fronts whose instantaneous position is well defined.
They transform fresh gas at temperature Tu ≈ 300 K into burnt gas at temperature
Tb ≈ 1200–3000 K.
For the case of a curved flame front propagating in a nonhomogeneous and/or nonsta-
tionary flow, the propagation speed is defined as the speed of the local normal to the front
with respect to the initial reactive mixture just ahead of the front. This notion is well defined
if the spatial and temporal scale of velocity variations in the upstream flow are much larger
than the thickness of the flame and the transit time of a fluid particle across the flame,
respectively. This is often the case. The propagation speed of a planar deflagration front in
a quiescent reactive mixture, called the laminar flame speed and designated by UL in the
rest of this book, is defined without ambiguity; see Fig. 1.1.
In automobile petrol engines, a flame is ignited by a spark plug and propagates as a
wrinkled premixed flame through the turbulent gas mixture confined in the cylinder. In a
Bunsen flame such as that of Fig. 1.2, the conical premixed flame is stationary in the frame
(a)
(b)
Figure 1.2 Bunsen flame. (a) Tomographic cut through a Bunsen flame. The unburnt gas is seeded
with microscopic refractory solid particles, which diffuse the laser light sheet used to visualise the
flow (courtesy of J.Quinard, IRPHE Marseilles). (b) The flow lines are strongly deflected through the
flame front; see also Fig. 2.6.
of the burner. The unburnt reactive mixture flows out of the orifice in a laminar flow with a
velocity greater than flame speed, UL , and the inclination of the flame spontaneously adjusts
itself so that the component of the gas flow normal to the flame front is equal to UL . Increas-
ing the flow rate will decrease the angle of the summit and increase the height of the cone.
For a planar flame, the product of the initial density of the unburnt gas and the laminar
flame speed, ρu UL , defines the mass of reactive mixture transformed into burnt products per
unit time and per unit surface of a planar flame, m = ρu UL . When the internal structure of
the flame is not significantly modified by wrinkling of the front, the mass flow rate remains
close to ρu UL and the total mass of gas burnt per unit time is given by ρu UL S, where S is
the total flame surface area. This is true when the wavelength of wrinkling is much greater
than the thickness of the flame. However, at the tip of a Bunsen cone the curvature is very
high and a particular study is required.
Non-Premixed Combustion
In non-premixed combustion, the fuel and the oxidant are separated by the reaction front,
fed by diffusive fluxes, and the flame is called a diffusion flame; see Fig. 1.3. This mode
of combustion does not propagate in the preceding sense, but propagation along the wall
of a solid combustible is possible. It is the type of flame found in a candle or a cigarette
lighter. In industrial gas-fired burners (furnace, boilers, ovens, etc.) the chemical reaction
develops in the mixing layer between the jet of fuel gas and the ambient air (oxidant); see
Fig. 1.4. In this book attention is focused on premixed systems, and diffusion flames will
not be covered. For more detail, readers are referred to classical books.[1,2,3] A brief outline
[1] Williams F., 1985, Combustion theory. Menlo Park, Calif.: Benjamin/Cummings, 2nd ed.
[2] Turns S., 2000, An introduction to combustion. McGraw-Hill, 2nd ed.
[3] Law C., 2006, Combustion physics. Cambridge University Press.
(a)
(b)
Figure 1.3 Photo of a candle flame in micro-gravity. In the absence of gravity-induced convection,
the reaction zone is blue and the flame is hemispheric. (b) Diffusive fluxes in the region of the flame.
The heat is evacuated in the same direction as the combustion products. Photo from nasa.gov website
with thanks.
Figure 1.4 Diffusion flame in a sheared mixing layer. The black arrows represent the flow of fuel
and oxidant separated by a wall, at the end of which the diffusion flame is anchored. Between pairs
of roll-ups, stagnation points appear similar to that shown in Fig. 1.7, where the flame is locally
stretched.
is presented in Section 1.2.4. The two types of flame (premixed and diffusion) can coexist
in highly turbulent combustion chambers, when fuel and oxidant are injected separately.
1.2.2 Premixed Combustion

In this section we recall the basic concepts in the theory of premixed combustion.[1]
Combustion Temperature and Equivalence Ratio

The energy liberated by ordinary combustion comes from the change in binding energy
between the atoms forming the molecules. The binding is due to interactions of the elec-
trons in the external layers of the atomic structure. The order of magnitude of this energy is
a few electron-volts (eV) per molecule, leading to a temperature increase of few thousand
Kelvin; see Section 14.1.1.
The temperature of combustion of premixed gases results from a chemical equilibrium;
see Section 14.2. This temperature varies with the composition of the mixture. It reaches
a maximum when the proportion of fuel and oxidant is near stoichiometry, defined as the
[1] Clavin P., 1994, Ann. Rev. Fluid Mech., 26, 321–352.
Figure 1.5 Calculated burnt gas temperature of methane–air flames as a function of equivalence ratio.
composition for which both are totally consumed, for example two moles of hydrogen
mixed with one mole of oxygen, 2H2 +O2 → 2H2 O. More generally, if ϑF+ moles of fuel F
fully react with ϑO+ moles of oxidant O to form combustion products P: ϑF+ F + ϑo+ O P,
then a fresh mixture containing NF moles of fuel and NO moles of oxidant is said to have a
stoichiometric composition if NF /NO = ϑF+ /ϑO+ .
The equivalence ratio φ of the mixture is defined by
NF /NO
φ= . (1.2.1)
ϑF+ /ϑO+
A stoichiometric mixture is characterised by φ = 1. In a rich mixture, φ > 1, the fuel is in
excess and the combustion is limited by the oxidant. For lean mixtures, φ < 1, the oxidant
is in excess. An example of the evolution of the combustion temperature of methane–air
flames with equivalence ratio is shown in Fig. 1.5.
Chemical Kinetics
The chemical kinetics controlling the reaction rate of the combustion of hydrogen and
common hydrocarbon fuels is complex. Nevertheless the system has overall global prop-
erties that can be reasonably well approximated by simple kinetic models, described in
Chapter 5. The overall chemical heat release rate increases very strongly with tempera-
ture, and the combustion rate undergoes a thermal runaway, which stops only when the
minority reactant has been completely consumed. The energy release rate is significant
only for temperatures typically above 1000 K. At room temperature a reactive mixture can
be conserved indefinitely: the composition of the mixture is totally frozen far from thermo-
chemical equilibrium. However, at high temperature, the characteristic reaction time (time
to reach thermo-chemical equilibrium) is short, of the order of a few microseconds at the
temperature produced by the exothermic reaction, typically 2000 K.
The thermal runaway in combustible mixtures can be described in macroscopic terms,
as a function of the temperature and pressure. It is not necessary to invoke Boltzmann’s
equation. This is because the frequency of inelastic collisions, which liberate chemical
energy, is much lower than the frequency of the elastic collisions, which maintain the
(Maxwell–Boltzmann) equilibrium distribution of velocity. In other words, a reactive
molecule undergoes many elastic collisions before being transformed. The liberation of
energy thus proceeds on a time scale that is long compared with the time needed to reach
local equilibrium, and the temperature remains locally and globally a well-defined quantity
during the chemical reaction.
Thermal Sensitivity and the Arrhenius Law

The characteristic reaction time, τr , of a first-order elementary reaction between two reac-
tive species, R1 + R2 → P1 + P2 , is defined in a homogeneous medium by the evolution
equation dn1 /dt = −n1 /τr , where n1 = N1 /V is the number density of the limiting reactive
species R1 . Usually, the reaction time τr is a function of the temperature, T, and proportional
to the density n2 of the species R2 . To simplify the presentation, we will limit the discussion
in this section to the situation where the species R2 is in great excess, such that the relative
quantity consumed by the reaction is very small. The change of concentration is negligible
in the expression for τr , which is then only a function of the temperature, T. For most
gaseous combustible mixtures, far from the flammability limits, and over a wide range of
temperature, the reaction rate is described by the Arrhenius law,
1/τr ≈ e−E/kB T /τcoll with E/kB T 1, (1.2.2)
where τcoll is the mean time between elastic collisions of species R1 with R2 , and E is an
activation energy that is systematically greater than the mean energy of thermal agitation,
kB T, typically by an order of magnitude. The activation energy E can be interpreted as
the energy that must be brought by the molecules into a collision in order to open bonds
and initiate the chemical reaction. The exponential factor in the Arrhenius law comes from
the Maxwell–Boltzmann distribution of velocities. It represents the fraction of the number
of collisions (per unit time) in which the relative energy carried by the two molecules is
greater than E. According to (1.2.2), the thermal sensitivity of the reaction rate increases
as the ratio of activation energy to mean thermal energy, E/kB T, increases: δτr /τr ≈
δτcoll /τcoll − (E/kB T)δT/T, where, according to (1.2.3), the sensitivity of the collision
rate, 1/τcoll , to temperature at constant pressure is only T 1/2 , δτcoll /τcoll ≈ −(1/2)δT/T,
and can be neglected,
δτr /τr ≈ −(E/kB T)δT/T,
since, typically, E/kB Tb ∼ 8–12, where Tb is the maximum temperature in the flame,
Tb ∼ 1200–3000 K. The collision frequency can easily be evaluated by the elementary
kinetic theory of gases as follows. The distance travelled by a molecule during a lapse of
√
time δt is aδt where a ∝ kB T/mR is the mean velocity of the molecules (of the order
of the speed of sound), a few hundred metres per second for normal conditions, and mR is
the mass of a molecule. The volume swept, aσo δt, is expressed in terms of the collision
cross section σo = 4π ro2 where ro is the range of action of inter-molecular forces (a
few angstroms, 10−10 m). Introducing n, the number density of particles, the number of
collisions experienced by the test particle is naσo δt, and the number of collisions per unit
time is naσo . The mean free path , namely the mean distance travelled by a molecule
between two collisions, is obtained by the ratio of the distance travelled to the number of
collisions,
= 1/(nσo ), 1/τcoll = naσo ; (1.2.3)
see Section 13.3. This leads to a temperature dependence of τcoll as T −1/2 at constant
pressure, p = nkB T.
The Arrhenius law (1.2.2) then describes a strong variation of the reaction rate with
temperature. For example, taking a typical value for E/kB Tb = 8, the Arrhenius factor
at the temperature of the burnt gas is e−E/kB Tb ≈ 3 × 10−4 . So, for a typical mean time
between elastic collisions of one nanosecond, τcoll ≈ 10−9 s, the characteristic reaction
time in the burnt gas is of the order of 1 μs, τr (Tb ) ≈ τcoll eE/kB Tb ≈ 3×10−6 s. However, at
room temperature Tu , Tb /Tu ≈ 8, E/kB Tu = 64, the characteristic reaction time τr (Tu ) ≈
τcoll eE/kB Tu is of the order of 6 × 1018 s, which is longer than the estimated age of the
universe ≈ 4 × 1017 s, or 13 × 109 years! The reaction rate given by the Arrhenius law is
thus completely negligible at room temperature. The high sensitivity of the reaction rate
to temperature makes premixed flames very different from the reaction–diffusion waves
encountered in other fields, as for example in biophysics; see Section 8.3.
Cut-Off Temperature, Flammability Limits

The Arrhenius law (1.2.2) is a drastic simplification that is nevertheless useful to analyse
the thermal propagation of flames and detonations. However, certain details of chemical
kinetics also play an essential role, as explained in Chapter 5. This is the case, for example,
when the limits of flammability or detonability are approached. These limits correspond
to brutal transitions in the energy release rate. The behaviour of flames at these limits can
be qualitatively represented in a simple but discontinuous model by introducing a cut-off
temperature T ∗ ≈ 1000 K; see (9.1.1) and (9.1.2). The reaction rate is given by (1.2.2)
for temperatures above cut-off, T > T ∗ , but drops to zero below T ∗ , corresponding to the
crossover of chain-branching and chain-breaking reactions; see Section 5.2.2.
Cool Flames
The chemical reaction never stops suddenly below T ∗ , but the nature of the oxidation
reaction changes completely. The combustion is incomplete; the initial reactants are not
transformed into stable species such as water vapour and carbon dioxide, which are the
major species produced at thermo-chemical equilibrium. Instead, the initial reactive species
are degraded to form relatively stable intermediate species, or more exactly metastable
species, such as peroxides, or aldehydes, such as formaldehyde. Only a small part of the
available chemical energy is released, and the characteristic reaction time increases rapidly
as the combustion temperature drops below T ∗ . For common hydrocarbon molecules, the
cut-off temperature is in the range 900–1300 K. The chemical reactions that take place at
low temperature, T < T ∗ , play no significant role in the propagation of normal flames.
Nevertheless, chemical waves, and even successions of waves (called ‘cool flames’ since
they release little energy) can propagate below T ∗ . They can be produced in special labora-
tory conditions.[1] The analytical study of such reaction–diffusion waves[2] show that they
have nothing to do with ordinary flames, and they will not be treated in this book.
Dimensional Analysis
Several mechanisms participate in the propagation of laminar combustion fronts: the
amount and rate of heat release, the transport of heat and mass, and also compressible
phenomena characterised by the speed of sound. In usual gaseous flames and detonations,
the radiative transfer of heat is weak and can be neglected compared with diffusive
transport. The physical mechanisms can then be characterised by the following quantities:
• The amount of chemical energy, qm , released per unit mass of the mixture. This energy is
of the order of the square of the sound speed in the burnt gas, a2b , which is typically four
to ten times larger than in the fresh mixture, (ab /au )2 ≈ Tb /Tu .
• The exothermic reaction rate, 1/τr , which is a strongly increasing function of the tem-
perature of the gas.
• The molecular and thermal diffusion coefficients, D, which all have the same order of
magnitude in a gas.
Using dimensional analysis, it is possible to construct different velocities from these param-
eters. It will be seen that the orders of magnitude for the propagation velocity of laminar
√ √
flames are D/τr (Tb ), and qm for detonations.
1.2.3 Flames in Premixed Gas

In ordinary premixed combustion, the initial thermodynamic state of the reactants can be
gaseous, liquid (such as nitromethane or nitroglycerin1 ) or solid, such as the propellants
used in missiles and rocket engines. The temperature of the combustion products Tb is
determined by the heat of reaction (the chemical energy that is released) and is generally
in the range 1200–3000 K for air-based combustion. Since the exothermic reaction can
proceed only if the temperature is sufficiently high, typically above 1000 K, the reaction
zone is almost systematically located in the gas phase.2 In the rest of this book we will
consider gaseous mixtures.
1 To be exact, nitromethane and nitroglycerin are not mixtures of reactants, but monopropellants, i.e. metastable molecules, which
decompose under the action of heat to release heat and combustion products.
2 One notable exception is the oxidation of solid carbon (charcoal).
[1] Lewis B., von Elbe G., 1961, Combustion flames and explosions of gases. Academic Press.
[2] Nicoli C., et al., 1990, In P. Gray, G. Nicolis, F. Barras, P. Borkmans, S. Scott, eds., Spatial inhomogeneities and transient
behavior in chemical kinetics, 317–334, Manchester University Press.
Laminar Flame Speed, Thickness and Transit Time

The propagation speed of a planar laminar flame in a premixed reactive medium at rest is
the result of a direct coupling between the reaction rate (1.2.2) and the transport of heat
and species. Roughly speaking, the chemical heat, which is released on the hot side of
the reacting zone, is transferred by conduction to warm up the adjacent cold mixture to a
temperature at which the exothermic reaction starts, and so on. According to the elementary
kinetic theory of gases, the molecular transport of heat and species is controlled by Fourier’s
law (15.1.28), Jq = −λ∇T, and Fick’s law (15.1.10), respectively, with coefficients for
thermal diffusion, DT ≡ λ/ρcp , and molecular diffusion Di that have the same order of
magnitude,
D ≈ a2 τcoll ≈ a ≈ 2 /τcoll (1.2.4)
(see (13.3.29)) where the thermal velocity of molecules v is of the order of the speed of
sound a. The collision frequency 1/τcoll , and the mean free path , have been evaluated
above; see (1.2.3). Using√the perfect gas law at constant pressure, nT ≈ cst., and remem-
bering that a varies as T, the diffusion coefficients increase with temperature as T 3/2
and nD as T 1/2 . However, to simplify the presentation, this weak temperature dependency
is often neglected in flames compared with the exponential factor in the Arrhenius law
(1.2.2). This is not the case for the ablation front in ICF, presented in Section 1.3.
The large value of the activation energy in the expression for the reaction rate (1.2.2) is
an essential characteristic of combustion. An important consequence is that the propagation
speed of deflagrations is very subsonic. Using dimensional analysis to construct a flame
speed, UL , from the reaction rate, τr , and the diffusion coefficient, D, the flame speed takes
√
the form UL ∝ D/τr . The order of magnitude of the Mach number of the flame speed,
M ≡ UL /a, can then be found using (1.2.2) and (1.2.4):

UL ∝ D/τr ⇒ M ≡ UL /a ∝ e−E/kB T 1. (1.2.5)
No matter which temperature (between the fresh gas temperature, Tu , and the burnt gas
temperature, Tb ) is used in (1.2.5), the Mach number is very small since E/kB Tb ≈ 8 in
usual mixtures. Then, as explained in Section 2.1.1 the relative pressure variation across
the flame is negligible in front of the thermal energy. Therefore, according to (15.2.3), the
temperatures of the fresh and burnt gases are related by a simple thermal balance
cp (Tb − Tu ) = qm , (1.2.6)
where cp is the specific heat per unit of mass at constant pressure. Since the reaction
proceeds at high temperature, the relevant temperature in (1.2.5) is Tb and the Mach number
of flame propagation is found to be of order 10−2 , yielding a flame velocity of the order of a
few metres per second. This value is somewhat greater than those measured experimentally,
between a few tens of centimetres per second and 1 m/s for typical flame speeds in air (it
reaches 10 m/s only in very energetic mixtures such as a stoichiometric mixture of hydrogen
or acetylene in pure oxygen). This difference, discussed in Section 2.1.2, will be explained
by the analytical study of the thermal propagation of flames in Section 8.2. This points
out the limitations of dimensional analysis when large nondimensional parameters such as
E/kB Tb are involved in the problem. The laminar flame thickness, dL , is the characteristic
length over which the gas temperature rises from Tu to Tb ; see Fig. 1.1. Dimensional
analysis using (1.2.2) and (1.2.4) shows that dL is proportional to the inverse of the pressure
and is much greater than the mean free path, = D/a, when the activation energy is large:

dL ∝ UL τr ≈ Dτr ≈ eE/kB T . (1.2.7)
For a typical stoichiometric methane–air flame this estimation yields a flame thickness of
the order of 10−5 m. For the same reason as above, this value is an order of magnitude
too small compared with experimentally measured values. The flame transit time, τL , the
characteristic time taken by the gas to traverse the flame structure, of the order of the ratio
of flame thickness to flame speed, is, according to (1.2.2), much larger than the elastic
collision,
τL ≡ dL /UL ≈ τr τcoll . (1.2.8)
These results confirm that, thanks to the large activation energy, the structure of flames can
be described in terms of macroscopic variables, without resorting to the kinetic Boltzmann
equation.
1.2.4 Diffusion Flames

The discussion is limited here to a general presentation.
Internal Structure
In diffusion flames the reactants are initially separated in space. They may be injected with
different phases, as for instance a drop of liquid fuel burning in air. However, when one of
the reactants is initially in a condensed phase, it is vaporised before reaching the reaction
zone, except in a few exceptional cases, such as the combustion of carbon (charcoal). The
high temperature reaction zone is thus generally situated in the gaseous phase. It is fed by
molecular diffusion of the fuel from one side and oxidant from the other side. The reaction
products and the heat of combustion are evacuated by diffusion towards both sides, as
indicated in Fig. 1.3.
Mixing of fuel and oxidant must take place on the molecular scale before the reaction
can proceed. When the characteristic time of the exothermic reaction is sufficiently short,
in a sense to be defined later, the reaction zone is thin and the concentration of reactants
is very small in this zone, where they coexist but are rapidly consumed. However, the
diffusive fluxes of reactants, and thus the reaction rate, can be high. The fluxes of fuel and
oxidant have quasi-stoichiometric proportions and both reactants are completely consumed.
The combustion temperature is then close to that of a stoichiometric premixed flame. Since
the temperature around the reaction zone is high, fuel and oxidant cannot coexist outside the
Figure 1.6 Sketch of profiles of temperature and concentration through a diffusion flame.
Figure 1.7 Flow field around a stagnation point.
reaction zone. The position of the reaction zone must then adjust itself so that the reaction
occurs with a stoichiometric composition; see Fig. 1.6. The diffusion time, τdiff = L2 /D,
is usually controlled by a geometrical length L, and D is the diffusion coefficient of the
reactants. For example in fuel-droplet combustion, L is the radius of the flame, of the same
order of magnitude as the radius of the droplet. This time is usually much longer than
the reaction time, τdiff τr (Tb ), τr (Tb ) ≈ 10−5 –10−6 s. The reaction zone is thus much
thinner than the external diffusion zone, as anticipated. Moreover, the reaction rate, being
limited by diffusion, is of the order of diffusion rate, 1/τdiff . For example, the combustion
time of a stationary drop of liquid hydrocarbon in a quiescent atmosphere is given (to a
numerical factor of order unity) by τd ∝ d2 /D, where d is the initial diameter of the droplet
and D is the molecular diffusion coefficient of oxygen in the gaseous mixture containing
the combustion products.
For a diffusion flame stabilised in a stagnation flow between two opposed jets of fuel and
oxidant as in Fig. 1.7, the two parameters characterising the external diffusion zones are the
stretch time, τs , equal to the inverse of the velocity gradient, and the diffusion coefficient,
D, of the reactants. The consumption of reactants per unit surface and per unit time is given
Figure 1.8 Sketch of the configuration studied by Burke and Schumann to represent a diffusion flame
at an orifice.
√
by the mass flux of reactants, whose order of magnitude is ρYR D/τs where YR is the
initial mass fraction of reactant in the jet of density ρ.
In the more general case, such as a jet of fuel entering a flow of oxidant, the problem is
complicated by the fact that the shape of the flame front is no longer determined by simple
geometrical constraints. Nevertheless, the problem can be simplified by supposing that the
diffusion coefficients of fuel and oxidant are equal and that the reaction is a single-step
irreversible reaction. With this simplification, the term containing the chemical consump-
tion (reaction rate) can be eliminated from the conservation equations by a introducing a
linear combination, Z, of the concentrations of fuel and oxidant. This combination Z, called
the mixture fraction , is a conserved scalar governed by a convection–diffusion equation. It
can be solved using the boundary conditions of the configuration. The first analysis of this
type was performed by Burke and Schumann[1] for the geometry of a uniform coaxial jet
sketched in Fig. 1.8. In the limit of infinitely fast chemistry, the reaction zone is infinitely
thin and on this surface the concentrations of fuel and oxidant are zero. The location of the
reaction zone is given by the location of the surface Z = 0.
The effect of finite rate chemistry on a diffusion flame stabilised between two opposed
jets (Fig. 1.7) or in a mixing layer (Fig. 1.4) can lead to local extinction through stretch of
the reaction front. The first analytic studies of stretched diffusion flames were performed
in 1974 by Liñan.[2]
Diffusion flames and premixed flames can coexist in the same flow. This is the case, for
example, in the region where a diffusion flame is anchored in the mixing layer of Fig. 1.4.
Molecular mixing takes place in the near wake of the plate separating the flows of fuel and
oxidant. A zone of premixed gases is thus present in the wake. The flame is then anchored
by a triple flame formed by a premixed flame, with lean and rich branches stabilised in
this nonuniform premixed flow, followed by a diffusion flame, which develops downstream
in the mixing layer.[3] These triple flames can also appear on the edges of holes produced
in a diffusion flame by local extinction due to stretch. Another type of diffusion flame
anchoring, purely thermal in nature, can exist where the flame is directly attached to the
trailing edge of a wall that is at a high temperature.[3]
[1] Burke S., Schumann T., 1928, Ind. Eng. Chem., 20(10), 998–1004.
[2] Liñan A., 1974, Acta Astronaut., 1(7-8), 1007–1039.
[3] Fernández-Tarrazo E., et al., 2006, Combust. Flame, 144(1-2), 261–276.
1.2.5 Gaseous Detonations

In premixed combustion, there is also a supersonic mode of propagation, called detonation.
The history of the discovery of this phenomenon in the nineteenth century was briefly
recalled in the introduction.
ZND Structure
The structure of a detonation, called ZND in honour of the works of Zeldovich, von
Neumann and Döring in the middle of the twentieth century, is composed of an inert
shock wave followed by and coupled to an exothermic reaction zone. This structure was
initially imagined half a century earlier by the chemist Paul Vieille.[4] As explained in
the complements, recalled in Part III, the thickness of shock waves is microscopic (see
Section 15.1.7), so that these waves can be considered as hydrodynamical discontinuities
of the flow field. Their formation is recalled in Section 15.3. Since the propagation speed
of the wave, D, is supersonic, of the order of 1800 m/s for detonations of hydrocarbon–
air mixtures at room temperature and pressure, the compression in the leading shock
wave raises the pressure, density and temperature to values sufficiently high to initiate the
chemical reaction. The values of these quantities, pN , ρN , TN , just downstream of the shock,
called the Neumann conditions, are expressed in terms of the propagation Mach number
M ≡ D/au by the Rankine–Hugoniot relations (4.2.14)–(4.2.17). For M of the order of
five, the Neumann conditions are typically pN /pu ≈ 30, ρN /ρu ≈ 5 and TN ≈ 1800 K. The
high pressure in the burnt gas behind the front causes detonations to be very destructive.
The study of the structure of planar detonation waves is presented in Section 4.2. The
ZND structure of a detonation can be understood from Fig. 1.9, which represents schemat-
ically a (very) long tube filled with a reactive gaseous mixture. One extremity of the tube
is closed by a piston advancing at a constant speed. Viscous effects and the existence of
boundary layers at the walls are neglected (slip conditions at the walls). If the piston is
started initially from rest, the upstream gas is put in motion by a compressible wave, which
soon becomes a shock wave; see Section 15.3.4. If the shock is strong enough, the chemical
reaction is initiated in the gas at the Neumann condition just behind the shock, and, under
certain conditions, the reaction zone remains attached to the shock. The complex shock–
reaction structure then propagates at a constant supersonic speed, D, into the fresh gas.
The conservation equations for mass, momentum and energy show that the existence of
a stationary solution for a planar wave driven by a piston at constant speed, called an
‘over-driven detonation’, is possible only if the piston speed exceeds a minimum value;
see Section 4.2.3.
Chapman–Jouguet Regime
The Mach number of the downstream flow relative to the leading shock increases, as
chemical energy is released, from its initial value MN < 1 of the Neumann state to a
value Mb ≤ 1 at the downstream edge of the reaction zone in the burnt gas. The latter
[4] Vieille P., 1900, C. R. Acad. Sci. Paris, 131, 413.
Figure 1.9 In the reference frame of the laboratory (top), or in other words when the fresh gas is at
rest, the burnt gas behind a detonation moves in the same direction as the front, with a speed equal
to that of the piston. In the reference frame of the detonation (bottom), the fresh gas enters the front
with a supersonic velocity, D, greater than the subsonic speed UN at which the compressed burnt
gas leaves the front. Because of compressible effects, the temperature of the reacting gas reaches a
maximum before the exothermic reaction goes to completion.
value, Mb , increases towards unity as the speed of the driving piston is reduced, whilst
the detonation velocity, D, decreases towards a lower bound DCJ . The limiting value
Mb = 1, D = DCJ , corresponds to a critical nonzero value of piston velocity below
which there is no steady solution to the conservation equations across the detonation front,
as explained in Section 4.2.3. This marginal condition with unit Mach number (relative
to the front) of the burnt gas leaving the reaction zone is called the Chapman–Jouguet
(CJ) detonation. It will be seen in Section 4.2.3 that this solution can also propagate as an
autonomous detonation, that is, without the support of a piston. The propagation speed,
DCJ , can be derived directly from the conservation equations for mass, momentum and
energy (15.1.45)–(15.1.48). These conservation equations were established at the end of
the nineteenth century, and the expression for DCJ was already known at that time; see,
for example, the reference to Michelson (1893) in the books of Shchelkin and Troshin[1]
and of Zeldovich and Kompaneets.[2] The expression for DCJ simplifies when the chemical
energy released in the reaction is large compared with the initial thermal energy of the flow.
This is the case for almost all usual reactive mixtures where qm /(cp Tu ) is of the order of six
to ten in detonations. The propagation Mach number of a CJ detonation can then be written
[1] Shchelkin K., Troshin Y., 1965, Gasdynamics of combustion. Baltimore, Md.: Mono Book Corp.
[2] Zeldovich Y., Kompaneets A., 1960, Theory of detonation. Academic Press.

cp qm
qm /cp Tu 1: MCJ ≈ 2 +1 , (1.2.9)
cv cp Tu
where cp and cv are the specific heats per unit mass, supposed constant to simplify the
presentation; see Equation (4.2.44). Typically the planar detonation wave is unstable, lead-
ing to galloping detonations and a cellular structure on the detonation front. The analytical
studies of these phenomena are recent and are presented in Chapter 4.
Detonation Thickness
The thickness of a CJ detonation is greater than that of a subsonic flame for the same
thermodynamic conditions. Its order of magnitude is given by the product of gas velocity
UN and the reaction time, τr (TN ), at the Neumann state, τr (TN ) = τcoll eE/kB TN , according
to the Arrhenius law (1.2.2). In the reference frame of the detonation, the velocity, UN , of
the gas leaving the shock and entering the reaction zone is a fraction α of the local speed
of sound, UN = aN /α, with typically α ≈ 5 for a CJ detonation. A dimensional analysis,
similar to that used in (1.2.5), then shows that diffusive transport is negligible (see (4.2.45)
for more details), and the estimated detonation thickness dN is
dN = UN τr (TN ) ≈ (aN /α)τcoll eE/kB TN . (1.2.10)
This is a rough evaluation of the detonation thickness, sufficient for our purpose here. A
more accurate evaluation is given in Chapter 4; see (4.3.14). According to (1.2.7), and more
precisely to the result from asymptotic analysis (2.1.9)–(2.1.11), the thickness of a flame
can be rewritten
dL ≈ ab (E/kB Tb )τcoll eE/(2kB Tb ) , (1.2.11)
where ab is the speed of sound in the burnt gas at temperature Tb . Expressions (1.2.10) and
(1.2.11) then yield

Tb
dL E ab − k ET − 12
≈α e Bb TN
. (1.2.12)
dN kB Tb aN
The burnt gas temperature of a deflagration is higher than the Neumann temperature of a
CJ detonation, Tb /TN > 1. The exponential term in (1.2.12) is thus much smaller than
the prefactor, which is typically 30. For instance, for a CJ detonation in a stoichiometric
methane–air mixture with ambient conditions, TN ≈ 1525 K, aN ≈ 750 m/s, α ≈ 5,
τcoll ≈ 10−9 s, E/kB TN ≈ 11, and the value of the exponential is ≈ 5 × 10−4 , leading to a
detonation thickness of the order of a centimetre. This is two orders of magnitude greater
than the thickness of a flame of the same mixture. According to (1.2.10), the thickness of
overdriven detonations is smaller since the Neumann temperature increases with the speed
of the shock.
1.3 Fronts and Thermal Waves in Extreme Conditions

Explosions in astrophysical and terrestrial contexts have some aspects in common[1] but
also many differences. In principle, flames, shocks and detonations can also be produced in
the extreme conditions of thermonuclear fusion, namely at high temperature ≈107 –1010 K
and high density, 106 kg/m3 in ICF and much higher densities in massive stars: 1010 kg/m3
in the burning silicon shell surrounding the iron core and ≈1017 kg/m3 at the end of core
collapse.[2,3]
1.3.1 Thermonuclear Fusion

In 1926 Eddington[4] postulated nuclear reactions as the mechanism powering the stars.
Nuclear fusion reactions have some analogies with ordinary combustion. Nucleons (pro-
tons, neutrons) are held together by a nuclear binding energy to form an atomic nucleus.
This is analogous to the way in which atoms are held together to form a molecule. However,
nuclear forces are quite different, in strength and nature, from interatomic forces. Never-
theless, in the same way that energy is released in a chemical reaction when atoms are
rearranged to form different molecules having a higher binding energy, energy is released
if nucleons are rearranged to form more stable atomic nuclei. Also, a barrier of energy has
to be crossed in both cases before the energy can be released.
The differences arise from the binding energy of nucleons, several million electron-
volts (MeV), six orders of magnitude higher than the electronic binding energy of atoms
in molecules, typically a few electron-volts (eV). To fix ideas, we recall that an energy of
1 eV per particle (1 eV ≈1.6×10−19 joule) corresponds to a temperature of the order of
11 600 K. The energy released in nuclear reactions is thus typically six orders of magnitude
(106 ) greater than in chemical reactions. In order to make two nuclei interact, it is necessary
to overcome the potential barrier created by the electrostatic repulsion of their positive
charges; see Section 14.3.1. This barrier, called Gamow energy, EG , is also typically of the
order of a million electron-volts, implying that the temperatures needed to initiate nuclear
fusion are typically larger than 107 K. The conditions for nuclear fusion thus involve
completely ionised plasmas, and nuclear reactions occur at extreme conditions that are
far from that of ordinary combustion. Another difference is that the distinction between
fuel and oxidant can be irrelevant in nuclear fusion: two identical nuclei can combine and
release energy, as in the sun, for example, where protons (hydrogen nuclei) combine in a
succession of reactions to form a helium nucleus.
Fig. 1.10 shows the mean binding energy per nucleon as a function of the number of
nucleons in a nucleus. A flat maximum is reached for 56 nucleons, corresponding to atomic
iron 56 Fe (26 protons and 30 neutrons). The number in front of the symbol denotes the
[1] Wheeler J., 2012, Philos. Trans. R. Soc. London Ser. A, 370, 774–799.
[2] Janka H.T., et al., 2007, Phys. Rep., 442, 38–74.
[3] Janka H., 2012, Annu. Rev. Nucl. Part. Sci., 62, 407–451.
[4] Eddington A., 1926, The internal constitution of stars. Cambridge University Press.
1.3 Fronts and Thermal Waves in Extreme Conditions 31
9 12 235
Average binding energy (MeV)

C 56 U
8 16 Fe
4
O
7 He 238
U
6 7
Li
6
5 Li
4
3
3
3
H
He
2
2
1 H
1
0 H
0 30 60 90 120 150 180 210 240 270
Number of nucleons in nucleus
Figure 1.10 Binding energy per nucleon as a function of the number of nucleons in a nucleus. The
nucleus of iron (56 Fe) has the highest binding energy.
atomic mass (number of nucleons). The proton, denoted p, is simply 1 H and the nucleus
of the stable helium molecule, following hydrogen in the Mendeleev’s classification (see
Figs. 14.1 and 14.2), is denoted 4 He (2 protons and 2 neutrons). Energy can thus be released
either by the fission of heavy nuclei such as uranium or by the fusion of light nuclei, such as
1 H, 3 He (2 protons and 1 neutron), lithium 6 and 7 (6 Li, 7 Li), carbon 12 C, or oxygen 16 O. It
is fairly clear from Fig. 1.10 that potentially much more energy is released by the successive
fusion of light nuclei than by fission of heavy nuclei. Both processes stop with iron.
Due to the large energy barrier EG the rate of fusion reaction is very sensitive to tem-
perature in a way somehow similar to the combustion rate on earth but for different time
and energy scales; see (1.2.2) and (14.3.1)–(14.3.3). For a temperature about 2 × 108 K,
appreciably higher than the minimum (≈107 K) required for hydrogen burning into helium
recalled in Section 14.3.2, the gradual fusion of several 4 He nuclei into 12 C and 16 O
develops. Nuclear burning of mixtures of carbon 12 C and oxygen 16 O will start at higher
ignition temperatures, about 6 × 108 K and 1.5 × 109 K, respectively, to form essentially
28 Si and 24 Mg. Finally the fusion of two 28 Si nuclei into 56 Fe occurs at an even higher
temperature, about 4 × 109 K. Most of these thermonuclear reactions proceed through a

complex network of chain reactions that are not necessary to describe here.
1.3.2 Stellar Evolution and Supernovae

Stars are born from clouds of gas (essentially hydrogen and helium) collapsing under
gravitational attraction. They undergo a sequence of changes during their lifetime. The
knowledge of how stars evolve is developed by observing different stars at various states
of their lifetime. Depending on their initial mass, M, the lifetime of stars ranges from few
million years for the most massive stars to a time much longer than the age of the universe
for the least massive ones. This is because the core temperature increases with the size of
the star and because of the high sensitivity of thermonuclear burning to temperature, so that
massive stars burn all their fuel much faster than small stars.
Stellar Structure
Most of the energy transfer inside stars is due to radiation and can be roughly modelled
by a Fourier law (15.1.28) with a diffusion coefficient of the same form as (1.2.4), where
a is the speed of light and the mean free path of photons. In its quasi-steady state, a
star is in hydrostatic and thermal equilibrium:[1,2] the gravitational force is balanced by
pressure, while the radiative losses are balanced by the energy released in nuclear reactions.
The inner core temperature (roughly evaluated with an adiabatic hydrostatic equilibrium)
increases with the mass of the star. For a mass above 8% of the mass of the sun (M >
0.08 M , M ≈ 2 × 1030 kg is the solar mass), the core temperature of a star becomes
sufficiently high to ignite nuclear burning of hydrogen into helium. In a star, the nuclear
combustion inside the hot plasma cannot take the form of a laminar flame because of the
strong Rayleigh–Taylor instability associated with the gravitational force; see (2.2.15).
The resulting unsteadiness could lead to a combustion regime similar to that of a well-
stirred reactor or, more likely, of a strongly turbulent flame in the well-stirred regime; see
Section 3.1.3. This question is open and one cannot do better than to use empirical models
of nuclear combustion in stars. In the sun the radius of the central region of nuclear burning
is evaluated at about 25% of the total radius.
Unicity and stability of the equilibrium state solution are subtle questions[2,3] that are
briefly discussed in Chapter 7. In a stable configuration, contraction (expansion) of the star
increases (decreases) the pressure and the temperature of the core. The regulation mecha-
nism is roughly explained as follows (see the discussion following Equation (7.2.10)):
• If the fusion reaction runs too fast, the core heats up, producing a higher pressure and
making the core expand against gravity. Expansion cools core, slowing the rate of fusion.
• If the reaction runs too slow, the core cools, leading to a lower pressure and making the
core contract. Contraction heats the core, increasing the reaction rate.
When the hydrogen has been consumed and its nuclear burning rate is no longer suffi-
cient to balance energy loss by radiation, the star contracts, the core temperature increases
and the burning of helium into carbon and oxygen can be ignited. For stars of intermediate
mass (less than about eight times the mass of the sun), when the helium is exhausted and
after sufficient radiative cooling, the core of the star can take the form of an electron-
degenerate white dwarf composed of carbon and oxygen[4] in which degenerate electrons
offer resistance to further compression. For more massive stars, M > 8 M , the nuclear
burning of carbon and oxygen is ignited, producing heavy-element cores. For a mass greater
than ≈15 M the inner temperature increases sufficiently to ignite further fusion reactions,
and the final core is composed essentially of 56 Fe (the maximum nuclear binding energy).
[1] Chandrasekhar S., 1967, An introduction to the study of stellar structure. Dover.
[2] Zeldovich Y., Novikov I., 1971, Stars and relativity. Dover.
[3] Kippenhahn R., Weigert A., 1994, Stellar structure and evolution. Springer-Verlag, 3rd ed.
[4] Wheeler J., 2012, Philos. Trans. R. Soc. London Ser. A, 370, 774–799.
Supernovae
It was suggested as early as 1934 that the origin of cosmic rays is the explosion of massive
stars at the end of their life cycle. More quantitative considerations were developed later.[5]
Stellar explosions, the sudden disruption of an entire star at the end of its lifetime, are
called supernovae. They are relatively rare events, occurring on average one to three times
per century in a galaxy of the size of the Milky Way. They are observed with powerful
telescopes and the collected data are essentially the intensity and the spectrum of the light
versus time. The luminosity is typically 1 to 10 × 109 times brighter than the sun, often as
bright as the host galaxy of the exploding star. More precisely, the rate of radiated energy
is approximately 1034 –1037 J/s. These catastrophic events produce typically 1043 J of total
radiant energy, and, according to recent evaluations, some 100 times more energy is carried
away in the flux of neutrinos, most of of which are liberated in a violent short burst (a few
tenths of a second) at the beginning of the explosion of massive stars (neutrinos are light
and very energetic particles that are not easy to detect because they interact weakly with
matter). In general the luminous emission reaches a peak in a few weeks, and then the
luminosity fades over the course of months or years, and remnants can be observed during
hundreds and thousands years after the explosion of the star.
Remnants are materials that are ejected by the explosion at very high velocity (as much
as few tenths of the speed of light) into the interstellar medium. This medium is a very low
density hydrogen gas with temperature ranging from 10 K to 104 K and number density
ranging from 106 molecules per cm3 to less than one per cm3 , to be compared with 1019 per
cm3 for a gas at ordinary conditions on earth. According to (1.2.3), the order of magnitude
of the mean free path in the interstellar medium ranges from 103 to 109 km, still relatively
small on the length scale of one light-year, ≈ 1013 km, but not on the scale of the size of a
star, typically few 103 km for a white dwarf and 107 km for the external envelope of a high-
mass star. The freely expanding stellar material acts as a spherical piston, and, far away
from the initial radius of the star, it takes the form of a strong blast wave (100 times faster
than the speed of sound) that compresses and heats the interstellar medium up to 107 –108 K.
The radius of remnants reaches typically 10 light-years. The theoretical study of the struc-
ture of these astrophysical blast waves has a long history, reported in the literature.[6] They
are more complicated than the blast waves observed on earth and will not be discussed in
this book.
Supernovae play an important role in the universe by enriching interstellar structures
with higher mass elements. They are considered as standard candles by astronomers, useful
for measuring distance (using the luminosity) and expansion velocity (using the Doppler
shift). Supernovae are classified into two groups based on the presence or not of Balmer
series hydrogen lines in their spectra at maximum brightness. Those without hydrogen lines
are classified as type I (SNI), and those with hydrogen lines are classified as type II (SNII).
Type I supernovae (SNI) must have burned out, or ejected, their external shell of hydrogen
[5] Colgate S., Johnson M., 1960, Phys. Rev. Lett., 5, 235–238.
[6] Truelove J., McKee C., 1999, Astrophys. J., 120, 299–326.
before exploding. Type I supernovae (SNI) are brighter than type II (SNII) and decay faster.
There are also different subgroups of SNI and SNII that are not worth discussing here.
Astrophysics is an observational science. The current understanding of supernovae is
poor. More data can now be obtained from modern astronomical instruments, especially
concerning the neutrino outburst at the very beginning of the explosion (just before
the emission of visible light). These data are needed to improve our understanding. This
could be the case during this century if, by chance, a sufficient number of supernovae are
observed in nearby galaxies, as was the case in 1987 for the famous SN 1987A in the Large
Magellanic Cloud, visible to naked eye (approximately 166 000 light-years from earth;
by comparison, the diameter of the Milky Way is 100 000 light-years). Few supernovae
have exploded nearby; the last one in our galaxy was observed by Kepler in 1604. Most
of the hundreds of the present-day observations concern old supernovae far beyond the
Milky Way.
Without the possibility to perform experiments on earth, the current understanding of
star explosion is based on sophisticated numerical simulations. Due to a huge range of
scales, direct numerical simulations cannot be performed and empirical subgrid scale mod-
els have to be introduced (turbulent nuclear combustion, transport phenomena). Because
of their sensitivity to the models that control the star progenitor (initial condition), the
numerical results are questionable. Despite the ever-increasing complexity of the advanced
models in nuclear physics[1,2,3] (and an ever-increasing rate of publication), the key mech-
anisms of star explosion have not yet been identified. The understanding has not pro-
gressed much during the last 50 years. Two types of star explosions are currently retained:
a sudden reignition of nuclear fusion in the form of a thermonuclear explosion, and a
gravitational collapse-induced explosion. The first is retained for less massive progenitors,
such as white dwarf stars (SNI), while the second concerns the iron core of massive stars
(SNII). Both mechanisms involve waves of the same nature as those presented in this book
for earth conditions (flames and/or detonations for SNI, shock waves for SNII) but with
scales of energy, length and time that differ by many orders of magnitude.
Thermonuclear Explosion (SNI)

It has been known since 1931 that the equilibrium between the pressure of a relativistic
electron-degenerate gas and the gravitational force is no longer stable when the mass
increases above the Chandrasekhar mass, ≈ 1.4 M , and a gravitational collapse occurs.[4]
This concerns only matter sufficiently dense that the highly degenerated electron gas is
ultrarelativistic[5] with an adiabatic index γ = 4/3; see (13.2.45). The hydrostatic equilib-
rium is stable for γ > 4/3; see Sections 7.2 and 7.3.
[1] Bethe H., 1990, Rev. Mod. Phys., 62(4), 801–866.

[3] Burrows A., 2013, Rev. Mod. Phys., 85, 245–261.
[4] Landau L., Lifchitz E., 1986, Fluid mechanics. Pergamon, 1st ed.
[5] Landau L., Lifchitz E., 1982, Statistical physics. Part I. Oxford: Pergamon Press, 3rd ed.
Carbon-oxygen white dwarfs are small dense stars, ρ ≈ 109 –1011 kg/m3 . Their radius
is few thousands kilometres and their mass cannot be larger than the Chandrasekhar mass
since the degenerate electron gas is ultrarelativistic at high electron density ne > 1036
electrons/m3 ; see Section 13.2.4. White dwarfs can accumulate material from a stellar
companion (binary star system) and acquire a mass near to the critical Chandrasekhar mass.
They then begin to contract and quickly explode. The nature of the explosion is subject to
debate. A first scenario is the following. As a result of contraction-induced heating, carbon
fusion is ignited at the centre and stops the contraction. The nuclear combustion then takes
the form of a turbulent flame that is assumed to transit quickly to detonation. Due to the
Rayleigh–Taylor instability, a flame that is ignited at the centre and propagates outwards
is expected to be strongly unstable. Unfortunately the deflagration-to-detonation transition
(DDT) in such conditions is an open question, as it is also the case on earth for flames
expanding freely in open space; see Section 4.3.5. Another scenario, more popular today,
is the spontaneous initiation of a detonation similar to the one described in Section 4.3.4
for ordinary combustion. None of these unsteady combustion processes of ignition and/or
transition to detonation evoked to explain SNI explosions is clear. There is not even a
consistent evaluation of the corresponding orders of magnitude of time and length scales.
However, detonations seem to reproduce the characteristics of the light emitted by ejecta
more correctly than flames. The structure of unsteady planar detonation waves, sustained
by the nuclear burning of homogeneous mixtures of carbon–oxygen at the temperature
and pressure encountered in SNI, has been investigated numerically.[6] The results yield
typical detonation velocities ≈ 104 km/s in agreement with (1.2.9). According to (1.2.10),
the detonation thickness varies strongly with the Neumann temperature because of the
thermal sensitivity of the reaction rate. For nuclear fusion–sustained detonations the effect
is spectacular: the thickness of the detonation varies by many orders of magnitude with the
location inside the star,[7] from few centimetres at high density, ≈ 1012 kg/m3 , to 104 km
at density below 1010 kg/m3 , the latter being larger than the typical size of a white dwarf.
As developed throughout this book, the initiation, quenching and the various mechanisms
of instability of flames and detonations, as well as the transition from flame to detona-
tion, are relatively well understood in terrestrial conditions, even though they still require
further investigation, especially for the latter topic. Combustion knowledge results from
long-standing confrontations of analytical studies of simplified models and well-controlled
experiments. The situation is different in astrophysics. The mechanisms of SNI thermonu-
clear explosion have not yet been deciphered. They will not be discussed in the rest of this
book.
Core-Collapse Scenario for Supernovae (SNII)
The current status of the core-collapse scenario for high-mass stars is the result of more
than half a century of numerical investigations. It can be summarised as follows.[3,8]
[6] Dominguez I., Khokhlov A., 2011, Astrophys. J., 730, 87–102.
[7] Khokhlov A., 1993, Astrophys. J., 419, 200–206.
[8] Woosley S., Janka H., 2005, https://arxiv.org/abs/astro-ph/0601261, 1–11.
High-mass stars pass through successive stages of fusion, starting with the lightest
elements, hydrogen into helium, and ending with the the heaviest ones involved in nuclear
fusion, namely silicon into iron. The reason for the successive character of the fusion steps,
H→He, He→ C & O, C→ Ne & Mg, ..., O & Mg→Si, Si→Fe, is that each succes-
sive step requires a higher temperature and density than the previous one: the first one,
H→He, becomes significant for temperature of a few 107 K and density of a few thousand
kg/m3 , while O & Mg→Si requires T ≈ 1.9 × 109 K, ρ ≈ 8.8 × 109 kg/m3 , and the last one,
Si→Fe, T ≈ 3.3 × 109 K, ρ ≈ 4.8 × 1010 kg/m3 .
The numerical results[1,2,3] highlight an important mechanism of nuclear physics in
the conditions of high-mass stars: the neutronisation of nuclear matter by electron cap-
ture by free protons or nuclei along with the emission of a neutrino. This production
of neutrinos is an endothermic process that absorbs ≈ 0.78 MeV per neutron (essentially
the difference between the rest mass of a neutron and the couple proton plus electron),
accompanied by the emission of a neutrino. There are other mechanisms leading to neutrino
production and emission. At temperatures above 109 K, the production of electron–positron
pairs by gamma rays increases rapidly (a positron is the antimatter version of an electron,
predicted by Dirac in 1928 and observed in experiments shortly after). At temperatures
above 1011 K, an electron–positron pair can annihilate with the production of a pair of
neutrinos instead of photons.[2] Neutrino production increases strongly with increasing
density and temperature. However, for densities below 1014 kg/m3 the neutrinos, unlike
photons, escape easily from the star and produce an important cooling that significantly
reduces the heating by gravitation-induced compression. For densities above 1015 kg/m3 ,
neutrinos are trapped and participate in the local thermal equilibrium.[2] These mechanisms
are difficult to evaluate quantitatively but play an important role in the energy balance and
in the equation of state of nuclear matter.
When the fuel of a given fusion step runs out, the release rate of nuclear energy becomes
insufficient to balance energy loss by radiation and/or neutrinos. The star then begins to
contract; temperature and density both increase so that the ashes can be eventually ignited.
The next fusion step then produces energy release at a sufficient rate to balance the losses
and stop the contraction. These successive steps proceed up to the last one, namely the
production of iron, if the mass of the star is sufficient to generate the required temperature
and density at its centre. Otherwise the chain of fusion steps is stopped to form, for example,
a carbon–oxygen white dwarf. Also, the reaction rate of each step is very different. For
example, in a 15-solar-mass star, the duration of the first one, H→He, is typically 11 million
years and decreases regularly in the next steps; it is only 2 million years for the second one,
2000 years for the third and only 2 weeks for the last one.[4]

The current view is as follows.[1,2,3,4,5] Before the explosion of a high-mass star, the
progenitor model has an onion-skin structure, delimited by an external shell of residual
hydrogen surrounding successive layers of increasingly heavier elements, at increasingly
higher temperature and the pressure. The main characteristic of a high-mass star at the
end of its lifetime is the existence of a dense central core (ρ 1011 kg/m3 ) of iron and
neighbouring elements. There is no further fusion step to counteract contraction. The mass
of the iron core increases because it is fed by the surrounding shell burning silicon into iron.
The electrons of the core are strongly degenerated and ultrarelativistic (γ = 4/3). This is
not the case in the outer layers (γ > 4/3) that are less dense and not subject to an eventual
gravitational collapse; see Section 7.3. For a constant adiabatic index γ = 4/3 the mass
of a spherical structure in stable equilibrium must be smaller than an upper bound, called
the Chandrasekhar mass; see Sections 7.2 and 7.3. A catastrophic event occurs when the
mass of the iron core reaches the Chandrasekhar mass. The core collapses suddenly with a
maximum infall velocity as high as few tenths of the speed of light. Its size decreases from
a few thousand kilometres to a few tens of kilometres in a tenth of a second, or even less.
The response time of the stable outer shells being much larger (hours), the collapsing core
is decoupled from the outer shells.[1]
Before this event, the mass of the γ = 4/3 core increases slowly and the onion-skin
structure of the progenitor star is in quasi-steady state. The response of the whole star to
the sudden core collapse deserves further study. To the best of our knowledge no stability
analysis of the whole star has been performed; only semi-phenomenological analyses are
available.[6] Nevertheless, the case with a uniform value of the polytropic index γ is more or
less well understood; see Sections 7.2 and 7.3. Analytical studies have been performed[7]
for the marginal case γ = 4/3 showing the existence and the stability of a homologous
solution for the collapsing core presented in Section 7.3.2.
During the collapse, the density of the iron core increases abruptly and reaches ρ ≈ 4–
5 × 1017 kg/m3 in a few tens of seconds, typically twice the density of atomic nuclei. This
dense and neutron-rich sphere emits a huge flux of neutrinos during a short time, releasing
an energy equivalent to about 10% of the rest mass (a few 1046 J) in less than a second. If
it does not collapse into a black hole, it eventually settles down as a 10-km-radius neutron
star. A strong neutrino burst is the signature of a catastrophic core collapse.
A first difficulty in the theoretical analyses is the inhomogeneity of the steady state.
Due to gravitation and to entropy increase through reactions and dissipative transfers of
energy, the thermodynamic conditions (temperature, density) and the properties of the
matter (polytropic index, γ , and equation of state) vary with the radius. Unfortunately, this
is also the case across the collapsing core, essentially because new phenomena of nuclear
physics appear as the density increases.
The greatest difficulty is to explain how the SNII explosion involving typically 1044 J of
kinetic energy can be produced and how it can eject such an important fraction of the mass
[5] Cooperstein J., Baron E., 1990, In A. Petschek, ed., Supernovae, chap. 9, 213–266, Springer-Verlag.
[6] Zeldovich Y., Novikov I., 1971, Stars and relativity. Dover.
[7] Goldreich P., Weber S., 1980, Astrophys. J., 238, 991–997.
of the star, with such a high velocity, into the interstellar medium. The situation is currently
not clear and the SNII explosion is not understood. According to recent analyses, nuclear
burning of the external layers might contribute at most to ≈ 10% of the blast energy, so that
core-collapse supernovae must be gravitationally powered.[1] In the current view, neutrinos
are thought to play an essential role through energy transfers, as is discussed a few lines
below.
It is worth first recalling the situation before advanced nuclear physics was introduced
into the model. In the 1950s and 1960s it was noticed,[2] in the simplest hydrodynamical
calculations of gravitational core collapse in spherical geometry, that a strong outwards-
propagating shock wave is quickly formed. The shock appears suddenly near to the sonic
point, enclosing about half the mass of the collapsing core, not far from the radius where
the infall velocity is maximum, delimiting roughly the inner region where the velocity
profile is homologous (infall velocity proportional to the radius); see Section 7.3. As a
result, behind the shock wave, the flow velocity of the shocked material is reversed and
flows outwards. The rest of the inner part of the core is suddenly halted and set almost at
rest. The material outside the shocked sphere still flows inwards at a supersonic velocity
and is both compressed and reversed when crossing the shock. This abrupt transition is
called a rebound. In some conditions, the numerical calculations showed that the mass flow
rate across the shock is sufficiently large for the shock to escape from the collapsing core.
It can then propagate through the external shells and reach the exterior of the star. This
scenario is called the prompt shock. As the external surface of the star is approached, the
density decreases to zero as a power law due to radiative losses[3] (the external shell is
transparent to radiation). According to an analytical solution[2] (self-similar solution of the
second kind), the temperature and the velocity of the shocked gas diverges when the shock
wave emerges at the surface of the star. This phenomenon is due to a divergence of the
energy per unit mass, and can explain the acceleration to a tremendously high speed of the
particles ejected into the interstellar medium.
There is no analytical solution of the rebound. In contrast to what is generally said
in the literature, the rebound is not necessarily related to incompressibility of matter at
high density. It is also produced when the compressibility does not change much with the
density; see Section 7.3. In contrast to simple waves in quasi-planar compressible flows
recalled in Sections 15.3.1–15.3.4, the quasi-spontaneous formation of the shock during
gravitational collapse in spherical geometry is poorly understood. Simple models with
a ‘pressure versus density’ law introduced into Euler’s equations deserve further study.
When the most advanced models of nuclear physics at high density, ρ > 1010 kg/m3 , and
temperature T > 109 K are introduced into the energy equation (insofar as the physics
is known in such extreme conditions), the numerical simulations do not reproduce the
explosion.[1,4,5] The rebound is still produced but the prompt shock scenario does not work

[2] Zeldovich Y., Raizer Y., 1967, Physics of shock waves and high-temperature hydrodynamic phenomena II. Academic Press.
[3] Zeldovich Y., Raizer Y., 1966, Physics of shock waves and high-temperature hydrodynamic phenomena I. Academic Press.
because the shock cannot escape from the collapsing core.[4] The strength of the shock
is substantially weakened by the copious energy loss due to both escaping neutrinos and
the dissociation of heavy nuclei[4] (mainly Fe) into nucleons (≈ 8.8 Mev per nucleon).
A few milliseconds after its formation the shock takes the form of an accretion surface of
discontinuity, delimiting a small sphere of very hot and dense material quasi at rest (a hot
neutron-rich core called a proto-neutron star) that is accreting mass at a few tenths of solar
mass per second. This inner dense part of the core is further compressed by gravitation,
emitting neutrinos. In the numerical simulations[1,6] the shock appears typically at about
10 km from the centre and its radius does not grow beyond 100–200 km.
The objective of current investigations is to find a mechanism that uses a small part of the
tremendous energy emitted in the form of neutrino flux, sufficient to turn the core collapse
and imploding matter into an explosion of the star. The main part of this neutrino flux is
emitted from the so-called neutrino sphere, located at density about 1014 kg/m3 , at a radius
smaller than the stalled accretion shock. The most popular mechanism for a neutrino-
driven explosion is absorption of neutrinos, heating the material. This requires a good
knowledge of the transport, emission, absorption and re-emission of neutrinos. Because of
the lack of experimental data to validate the theoretical predictions in nuclear physics, the
corresponding laws are not well known. In spherical geometry, two ways to take advantage
of a hypothetical neutrino-heating mechanism have been envisaged, either through a local
revival of the shock wave[5] or through a parametric analysis of the nonexistence and/or
of the global stability of the quasi-steady state solution involving the stalled accretion
shock and the flow modified by the flux of neutrinos.[1,7] Other ideas investigated concern
multidimensional instabilities. Modest successes have been obtained so far.
Back to Gravitational Collapse

After 40 years of intensive investigation, the current status of the theory of core-collapse
explosion of supernovae is somewhat confusing. Knowledge could be improved by follow-
ing an approach similar to that successful in the theory of combustion. In the same way as
for the transient regimes of ordinary flames and detonations studied in this book, a drastic
simplification of the complexity might be useful (necessary?) to decipher the explosion
of stars. In that respect, parametric studies of simplified models, identifying the critical
parameters for the transition from ‘prompt shock to stalled shock’ would be welcome. An
eventual transition to explosion in the onion-skin structure of the external shells, subsequent
to a sudden core collapse, could also be investigated without invoking a prompt shock
scenario. Attention could also be paid to an eventual transition in the combustion regime
of silicon into iron (DDT?). Even though the nuclear energy available is too small (smaller
than the kinetic energy of the ejecta) to postulate such phenomena as the only mechanisms
exploding the star, they could play an important role in the whole process.
The difficult problems in nuclear physics involving neutrinos will not be addressed
in this book, in which only simple models are considered. The classical hydrodynamical

[7] Keshet U., Balberg S., 2012, Phys. Rev. Lett., 108, 251101.
results, at the root of gravitational collapse, which have been well established for a long
time, are presented in Sections 7.1 and 7.2. More recent results concerning the dynamics are
presented in Section 7.3. The physics of ultrarelativistic degenerated electrons, responsible
for the critical value 4/3 of the adiabatic index γ , is recalled in Section 13.2.4. An
illustration of the sensitivity of the evolution of the shock wave to the equation of state[1]
is given in Section 7.3.3 where a nonintuitive result is presented with an ultrasimplified
model: the so-called rebound and the prompt shock scenario are easily produced by a soft
equation of state having only a small variation of the compressibility with density. In
contrast, the prompt shock scenario is not observed with the rebound produced by a stiff
equation of state in which the compressibility of the dense material (the nuclear matter
at the centre of the core) is negligible compared with that of the less dense material at
the periphery. Moreover a mechanism of neutrino-driven explosion, based on a thermo-
acoustic instability, is suggested in Section 7.3.3.
1.3.3 Inertial Confinement Fusion

Another example of thermal waves in extreme conditions is the ablation front in ICF.[2] The
notion of ablation front relies on the steepness of the temperature gradient created by the
nonlinear temperature dependence of electron heat conductivity in fully ionised plasma.[3]
The control of the hydrodynamic instabilities of the ablation front is a major challenge
facing ICF.[2]
Context
In January 1951 Ulam and Teller developed the principle of a radiation-driven nuclear
explosion. On 31 October 1952 a successful test of a nuclear weapon (H-bomb) with
a yield of 10.4 megatons (of high explosive) was carried out on Enewetak Atoll. With
the development of pulsed high-power lasers after the 1960s (10 kJ with a pulse duration
of few nanoseconds at 1 μm wavelength in 1980), the idea of laser-driven implosion of
small cryogenic fuel spheres filled with a mixture of deuterium and tritium (D-T) was
developed with the objective of achieving the necessary 103 - to 104 -fold compression and
the temperature of a few 107 K required to ignite a nuclear fusion reaction. A brief summary
of the D-T reaction is recalled in Section 14.3.3. In this process, called inertial confinement
fusion (ICF), the mechanism to confine the matter is mass inertia. The initial objective
was the production of (nonweapon) energy through micro-explosions. However, the more
recent programs in France and the United States are more weapons oriented. In 2014 the
focus of ICF experiments is on understanding thermonuclear ignition under the conditions
found within the nuclear weapons, rather than production of energy. The objective is mainly
to provide experimental data for the validation of numerical simulation codes that are
used in the development of newer generations of nuclear weapons. The National Ignition
[1] Swesty F., et al., 1994, Astrophys. J., 425, 195–204.

[2] Atzeni S., Meyer-Ter-Vehn J., 2004, The physics of inertial fusion. Clarendon Press–Oxford Science Publications, 1st ed.
[3] Spitzer L.J., 1962, Physics of fully ionized plasmas. New York: Wiley Interscience, 2nd ed.
Facility (NIF), completed in 2009 in Livermore, California, aims to deliver a peak flash of
ultraviolet light of more than 106 J and 1014 W of energy and power into the target chamber.
The total amount of energy deposited in the fuel is a few 105 J and the laser is driven by
more than 108 J of electricity! In France an ICF facility, the Laser Megajoule (LMJ), built
near Bordeaux and equivalent to the NIF, has been officially operational since 2014. Both
the NIF and the LMJ use indirect drive. In this approach the target is surrounded by a metal
cylinder which is irradiated by the laser beams that are focused inside the cylinder. The
inner side of the cylinder is heated to a hot plasma, which radiates X-rays that are then
absorbed by the target surface. This is more efficient than direct absorption of laser light by
the small fuel sphere. At the end of 2013, experiments at the NIF are reported showing a
production of typically 5 × 1015 neutrons generated per shot. The nuclear energy released
is slightly more than the laser energy absorbed by the target. However, this is still a long
way from ignition.
Conditions for Ignition

For obvious reasons, the nuclear energy released from a single micro-explosion in a lab-
oratory experiment has to be limited to typically a few 108 J, representing the complete
burn of no more than 10−3 g of a D-T mixture. This corresponds to the energy released
by approximately 80 kg of high explosive, but the blast of such a nuclear explosion is
considerably weaker because the neutrons and photons carry relatively little momentum
per unit energy. An extreme level of compression has to be achieved in order to ignite such
a small D-T mass. As a minimum, the mean free path, defined in (1.2.3), must be much
smaller than the size of the fuel sample R. In reality the criterion is more elaborate because
the confinement time in ICF is very short, typically a few 10−10 s, and the reaction, which
is expected to be ignited at the centre, must have sufficient time to propagate throughout
the compressed fuel. A rough estimate is obtained by comparing the reaction rate (14.3.2),
n σnr v, σnr v ≈ 5 × 10−16 cm3 /s at T = 5 × 108 K, and the strong expansion rate that
develops immediately after the maximum of compression. The later can be evaluated as
a/R, where R is the size of the compressed fuel sample and a the speed of sound. Requiring
that the reaction rate be larger than the expansion rate yields a lower bound for the quantity
nR, where n is the number density. For the maximum cross section of the D-T nuclear
reaction shown in Fig. 14.5, namely for T ≈ 108 K, this imposes that ρR be greater than
few g/cm2 , where ρ is the fuel density. For 1 mg of D-T this corresponds to a density of
a few 102 g/cm3 , which is typically a density ratio of 103 relative to the initial density
of the cryogenic D-T sample in solid state (ρ ≈ 0.2 g/cm3 ). Such a high compression is
achieved by ICF; see below. It is worth stressing that, as illustrated by quasi-isobaric flame
ignition in Sections 2.4.2–2.4.5 and by the initiation mechanisms of detonations in Section
4.3, the full determination of the conditions of ignition in ICF is a difficult problem. It is
not addressed in this book. Attention will be limited to the hydrodynamical instability of
the ablation front during the implosion of the irradiated capsule. Various other scientific
problems involved in ICF are discussed in the specialised literature.[2,4]
[4] Lindl J., 1998, Inertial confinement fusion. Springer.
Figure 1.11 The principles of inertial confinement fusion.
Orders of Magnitude in Direct Drive

The order of magnitude of parameters and performances that are expected for a direct-
drive laser target[1,2] are the following. The target consists of a hollow shell capsule of
1.7×10−3 g cryogenic D-T mixture initially in solid state (T ≈ 17.9 K, ρ ≈ 0.22 g/cm3
at atmospheric pressure). The radius is slightly less than 2 mm and the thickness is typi-
cally 0.3 mm. The capsule is filled with D-T vapour (ρ ≈ 0.3×10−3 g/cm3 at atmospheric
pressure) and is coated with an outer plastic layer of thickness slightly less than 0.4 mm;
see Fig. 1.11. Irradiation leads to surface ablation; the plastic layer is vaporised, ionised
and expelled outwards. Due to momentum conservation, the capsule moves inwards at
high velocity (≈ 4×105 m/s) under the ablation pressure (≈ 108 bar). The fuel of the
imploding shell comes suddenly to rest near the centre. The pressure and density suddenly
increase to a few 1011 bar and ρ ≈ 400 g/cm3 , while the temperature of the compressed
D-T mixture inside the shell reaches about 10 keV (≈ 108 K). Ignition occurs in the central
hot spot and propagates outwards into the compressed fuel of the shell in the form of a very
fast thermal wave (5×106 m/s), driven by electron heat conduction and the escaping α-
particles. The fusion energy, typically few 108 J, is liberated in less than 10−10 s, before the
subsequent expansion of the hot matter quenches the nuclear reaction. Inertial confinement
is the best method on earth to reach the high density by compression (density ratio) that is
required for thermonuclear ignition. By comparison, the compression across a planar shock
front in a gas is limited to a factor typically between 4 and 6 in an ideal gas, ρN /ρu <
(γ + 1)/(γ − 1); see (4.2.14). Moreover, according to the first law of thermodynamics
(13.1.3), de = Tds − pd(1/ρ); isentropic compression (ds = 0) minimises the mechanical
energy −pd(1/ρ) > 0 needed to produce a given increase of internal energy de. It follows
that the formation of strong shocks should be avoided and temporal shaping of the driving
pulse helps to optimise a quasi-isentropic compression. The spherical shape of the capsule
is optimal to reach an implosion velocity sufficiently high to achieve the high density
[1] Atzeni S., Meyer-Ter-Vehn J., 2004, The physics of inertial fusion. Clarendon Press–Oxford Science Publications, 1st ed.
[2] Bychkov V., et al., 2015, Prog. Energy Combust. Sci., 47, 32–59.
1.4 Appendix: Physical Constants and Conversion of Units 43
required for ignition. The high temperature is attained through compression of the D-T
mixture inside the imploding shell.
Hydrodynamic Instability
At the peak irradiation density (typically 1015 W/cm2 ), the ablation pressure and the implo-
sion velocity reach 108 bar and a few 105 m/s in few tens of nanoseconds. Such a large
acceleration produces a very strong Rayleigh–Taylor instability (2.2.14) and (2.2.15) of the
ablation front since the density of the vaporised matter is much smaller than that of the
plastic layer. This tends to deform the imploding spherical shell and hinders the formation
of the central hot spot by breaking the spherical symmetry. One of the major challenges fac-
ing ICF is to control this instability. Fortunately, heat transfer by electron heat conduction,
involved in the ablation mechanism, stabilises the small wavelength disturbances reducing
the linear growth substantially. This scientific problem has a long history, starting in the
1970s.[3,4,5,6] A better understanding and appropriate description of the hydrodynamic
instability of the ablation front in ICF were developed later by comparison with flame
theory,[2,7,8,9] ending up with a simple model equation to describe the nonlinear dynamics
of the unstable ablation front.[10] This topic is presented in Chapter 6, and the details of the
analyses are given in Chapter 11.
1.4 Appendix: Physical Constants and Conversion of Units
G = 6.674 08 × 10−11 N m2 /kg2 Gravitational constant

c = 2.997 92 × 108 m/sec Velocity of light in a vacuum
−34
h̄ = h/2π = 1.054 57 × 10 J sec Planck’s constant
−23
kB = 1.380 65 × 10 J/K Boltzmann’s constant
N = 6.022 14 × 10 /mole 23
Avogadro’s number
qe = 1.602 18 × 10−19 Coulomb Electronic charge
−31
me = 9.109 38 × 10 kg Electron rest mass
−27
mn = 1.674 93 × 10 kg Neutron rest mass
[3] Bodner S., 1974, Phys. Rev. Lett., 33, 761–764.

[4] Takabe H., et al., 1983, Phys. Fluids, 26(2299-2307).
[5] Sanz J., 1994, Phys. Rev. Lett., 73, 2700–2703.
[6] Bychkov V., et al., 1994, Phys. Plasmas, 1, 2976–2986.
[7] Clavin P., Masse L., 2004, Phys. Plasmas, 11, 690–705.
[8] Clavin P., Almarcha C., 2005, C. R. Mécanique, 333(379-388).
[9] Sanz J., et al., 2006, Phys. Plasmas, 13, 102702.
[10] Almarcha C., et al., 2007, J. Fluid Mech., 579, 481–492.
mp = 1.672 62 × 10−27 kg Proton rest mass

−11
aB = 5.291 77 × 10 m Bohr radius (h̄2 /me qe )
1 bar = 105 Pa = 106 erg/cm3

1 J = 107 erg
1 eV = 1.602 18 × 10−19 J
1 eV/kB = 11 604.5 K

General Considerations: Nomenclature

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

General Considerations: Nomenclature

Uploaded by

Copyright:

Available Formats

1

Nondimensional Quantities and Abbreviations

Superscripts, Subscripts and Math Accents

1.1 Introductory Remarks

1.2 Combustion Waves on Earth

1.2.2 Premixed Combustion

Combustion Temperature and Equivalence Ratio

Thermal Sensitivity and the Arrhenius Law

Cut-Off Temperature, Flammability Limits

1.2.3 Flames in Premixed Gas

Laminar Flame Speed, Thickness and Transit Time

D ≈ a2 τcoll ≈ a ≈ 2 /τcoll (1.2.4)

τL ≡ dL /UL ≈ τr  τcoll . (1.2.8)

1.2.4 Diffusion Flames

Figure 1.7 Flow field around a stagnation point.

1.2.5 Gaseous Detonations

[4] Vieille P., 1900, C. R. Acad. Sci. Paris, 131, 413.

dN = UN τr (TN ) ≈ (aN /α)τcoll eE/kB TN . (1.2.10)

dL ≈ ab (E/kB Tb )τcoll eE/(2kB Tb ) , (1.2.11)

1.3 Fronts and Thermal Waves in Extreme Conditions

1.3.1 Thermonuclear Fusion

Average binding energy (MeV)

temperature, about 4 × 109 K. Most of these thermonuclear reactions proceed through a

1.3.2 Stellar Evolution and Supernovae

Thermonuclear Explosion (SNI)

[1] Bethe H., 1990, Rev. Mod. Phys., 62(4), 801–866.

[1] Burrows A., 2013, Rev. Mod. Phys., 85, 245–261.

[1] Burrows A., 2013, Rev. Mod. Phys., 85, 245–261.

Back to Gravitational Collapse

[6] Woosley S., Janka H., 2005, https://arxiv.org/abs/astro-ph/0601261, 1–11.

1.3.3 Inertial Confinement Fusion

[1] Swesty F., et al., 1994, Astrophys. J., 425, 195–204.

Conditions for Ignition

[4] Lindl J., 1998, Inertial confinement fusion. Springer.

Figure 1.11 The principles of inertial confinement fusion.

Orders of Magnitude in Direct Drive

1.4 Appendix: Physical Constants and Conversion of Units

G = 6.674 08 × 10−11 N m2 /kg2 Gravitational constant

[3] Bodner S., 1974, Phys. Rev. Lett., 33, 761–764.

mp = 1.672 62 × 10−27 kg Proton rest mass

1 bar = 105 Pa = 106 erg/cm3

You might also like

τL ≡ dL /UL ≈ τr τcoll . (1.2.8)