Professional Documents
Culture Documents
Frisch RadTransfer
Frisch RadTransfer
Frisch RadTransfer
Hélène Frisch
Radiative Transfer
An Introduction to Exact and Asymptotic
Methods
Hélène Frisch
Observatoire de la Côte d’Azur, CNRS,
Laboratoire Lagrange
Université Côte d’Azur
Nice Cedex 04, France
© The Editor(s) (if applicable) and The Author(s), under exclusive license to Springer Nature Switzerland
AG 2022
This work is subject to copyright. All rights are solely and exclusively licensed by the Publisher, whether
the whole or part of the material is concerned, specifically the rights of translation, reprinting, reuse
of illustrations, recitation, broadcasting, reproduction on microfilms or in any other physical way, and
transmission or information storage and retrieval, electronic adaptation, computer software, or by similar
or dissimilar methodology now known or hereafter developed.
The use of general descriptive names, registered names, trademarks, service marks, etc. in this publication
does not imply, even in the absence of a specific statement, that such names are exempt from the relevant
protective laws and regulations and therefore free for general use.
The publisher, the authors and the editors are safe to assume that the advice and information in this book
are believed to be true and accurate at the date of publication. Neither the publisher nor the authors or
the editors give a warranty, expressed or implied, with respect to the material contained herein or for any
errors or omissions that may have been made. The publisher remains neutral with regard to jurisdictional
claims in published maps and institutional affiliations.
This Springer imprint is published by the registered company Springer Nature Switzerland AG
The registered company address is: Gewerbestrasse 11, 6330 Cham, Switzerland
To Uriel, Anne, and Thomas
Preface
vii
viii Preface
constructed by (Wiener and Hopf 1931) for the Milne problem, which describes the
temperature profile of a stellar atmosphere in radiative equilibrium. Their method
of solution, which now carries their name, is one the greatest achievements in
mathematical physics in the first half of this century. It relies on the properties of
analytic functions in the complex plane. In contrast, when the question is posed to
find the distribution of photons that have been reflected or transmitted through a
scattering medium of finite extension, a slab for example, then there is no explicit
expression for the radiation field, inside or outside the medium. However, there
are many interesting exact relations and integral equations, which can be solved
numerically. This topic is only briefly mentioned in this book.
Following the work of Wiener and Hopf, several methods have been developed
to tackle semi-infinite medium problems. One of the best known method in
astrophysics is the construction of a nonlinear integral equation for the famous H -
function, a sort of special function for half-space transport problems. Proposed for
the first time by Ambartsumian (1942), it largely avoids complex plane analysis and
relies on invariance properties of the radiative transfer process. Plasma physicists
became strongly interested in the subject in relation to the transport of neutrons
in nuclear reactors. New methods were developed, such as the Case singular
eigenfunction expansion method (Case 1960), which leads to linear singular integral
equations with Cauchy-type kernels. Methods of solutions for these equations,
also based on complex plane analysis, rely largely on a technique introduced by
Carlemann (1922) and developed by Muskhelishvili (1953). These methods lead
to boundary value problems in the complex plane, known as Riemann–Hilbert
problems, and provide explicit expressions for the H -function, as does the Wiener–
Hopf method.
A key motivation for writing this book was to present in the same volume several
of the methods leading to exact results for semi-infinite media, usually presented
in separate books or articles, for both scalar and polarized radiation fields, and also
connections, that are not always clearly apparent, between the various methods.
When the number of scatterings in the host medium is very large, asymptotic
techniques may be employed to analyze the large-scale behavior of the random
walk of the photons and to explain, for example, why monochromatic scattering has
all the characteristics of an ordinary diffusion process, while complete frequency
redistribution of spectral lines has in contrast those of a Lévy walk. This asymptotic
approach is presented in Part III of the book, while exact methods are presented in
Parts I and II. Part I deals with unpolarized radiation and Part II with monochro-
matic radiation, polarized by Rayleigh scattering, and spectral lines, polarized by
resonance scattering and the Hanle effect, which is a modification of resonance
polarization by a weak magnetic field. Details on the organization of each part are
given in Chap. 1.
The intended readership for the book ranges from first-year graduate students
to professional scientists. Astrophysicists can find in this book exact methods of
solutions used in radiative transfer, but also applied in distant fields, such as financial
mathematics. A major effort has been made in the organization of the material
to give a synthetic view on exact and asymptotic methods in radiative transfer.
Preface ix
Although oriented towards methods of solutions, this book also provides exact
expressions for the radiation field intensity and polarization for a number of standard
problems.
I am deeply grateful to many colleagues for their help and encouragement in
this undertaking. My sincere thanks go particularly to V. Bommier, V. V. Ivanov,
B. Rutily, and P. Zweifel, who have read significant parts of the book and have
taken the time to share with me their intimate knowledge of the field. I have adopted
many of their suggestions. My thanks go also to my family, especially to my mother
D. Piron-Lévy, for lasting encouragements; to my husband, Uriel, for reading the
full manuscript and playing the role of the non-specialist; and to L. Anusha, E. Lega,
and M. Sampoorna for their very generous help in the preparation of the figures. This
book would not exist without the continuous support of Springer teams in Europe
and India. I owe them a thousands thanks.
References
xi
xii Contents
Fig. 2.1 Left panel: the space variable τ and the inclination angle
θ of a ray with direction n. The axis z is normal to the
surface of the medium. Right panel: the heliocentric
interpretation of the angle θ . The Sun is viewed from the
right. Observations at disk center correspond to θ = 0 and
those at the limb to θ = π/2 . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 14
Fig. 2.2 The Doppler profile, and several Voigt profiles with
different values of the parameter a. The left panel shows
ϕ(x) and the right panel log ϕ(x). In the right panel, the
Lorentzian wing regime characterized by ϕ(x) ∼ 1/x 2
can be observed for |x| larger than (− ln a)1/2 .. . . . . . . . . . . . . . . . . . . . 21
Fig. 2.3 Complete frequency redistribution. The integration
domains (ξ, x) and (x, ξ ), with ξ = μ/ϕ(x) . . .. . . . . . . . . . . . . . . . . . . . 29
Fig. A.1 Integration contour for the Plemelj formulae. The
contribution from the singular point is obtained by letting
ρ → 0 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 65
Fig. 5.1 The function g(ξ ) for the Doppler profile and the Lorentz
profile, 1/(π(1 + x 2 )), in linear scales in the upper panel
and in log-log scales in the lower panel. The value of
g(ξ ) is constant for ξ ≤ 1/ϕ(0). The algebraic behavior
of g(ξ ) for ξ → ∞ given in Eqs. (5.8) and (5.9) can be
observed in the lower panel. The Voigt profile has the
same algebraic behavior as the Lorentz profile .. . . . . . . . . . . . . . . . . . . . 71
Fig. 5.2 The function k(ν) for the Doppler profile and the Lorentz
profile, 1/(π(1 + x 2 )), in linear scales in the upper panel
and in log-log scales in the lower panel. The angular
point is located at ν = ϕ(0). For the Doppler and Lorentz
profiles, k(ν) tends to zero as 1/ν for ν → ∞. For
ν → 0, one can observe in the lower panel the algebraic
behaviors given in Eqs. (5.8) and (5.9). The Voigt profile
has the same algebraic behaviors as the Lorentz profile . . . . . . . . . . . 72
xxi
xxii List of Figures
Fig. 12.3 The figure shows the strips of analyticity of V (z), the
dispersion function, and of V ∗ (z) defined in Eq. (12.34).
It shows also the analyticity half-planes of Vu (z) and
Vl (z), which satisfy V (z) = Vu (z)/Vl (z), and of Vu∗ (z)
and Vl∗ (z), which satisfy V ∗ (z) = Vu∗ (z)/Vl∗ (z) . . . . . . . . . . . . . . . . . . 225
Fig. 12.4 Analyticity half-planes of the inhomogeneous term
Q̂∗ (z) and of the functions Gu (z) and Gl (z) defined in
Eq. (12.47). The thick solid line indicates the branch cut
of Vl (z) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 229
Fig. 12.5 The analyticity half-planes of the functions in the left
and right hand-sides of Eq. (12.16) and their common
analyticity strip . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 230
Fig. 12.6 The Wiener–Hopf method for the Milne problem. The
half-plane analyticity domains of the functions in the left
and right hand-sides of Eq. (12.73) and their common
strip of analyticity. The function Vu (z) has a double zero
at the origin . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 235
Fig. 12.7 The Sommerfeld diffraction problem: (a) the geometry;
(b) the boundary conditions for the field ϕ(x, y) and
for ϕ (x, 0± ) = ∂ϕ(x, 0± )/∂y, the normal derivative of
ϕ(x, y) on the axis y = 0. The notation 0± means that
y → 0 by positive or negative values. The continuity
across the axis y = 0 is indicated by a vertical line with
two arrows . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 241
Fig. 12.8 The Sommerfeld diffraction problem: the factorization of
γ = (k − λ)1/2 (k + λ)1/2 . The principal determination of
(k + λ)1/2 is analytic in (k) > 0 and that of (k − λ)1/2
in (k) < 0 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 245
Fig. I.1 Monochromatic scattering. (a) Contours for obtaining an
explicit expression of Vu (z) in terms of the logarithm
of the dispersion function. The point z lies in the upper
half-plane (z) > δ > −ν0 . (b) Contour for obtaining the
relation between Vu (z) and X(z). The point z lies in the
upper half-plane (z) > −1 . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 247
Fig. 13.1 Ellipse described by the tip of the electric field vector in
the plane orthogonal to the direction of propagation. The
angle (0 ≤ ≤ π) specifies the orientation of the
ellipse and the angle χ (−π/4 ≤ χ ≤ π/4) the ellipticity
and the sense in which the ellipse is being described . . . . . . . . . . . . . . 257
Fig. 13.2 Rotation of the reference direction by an angle α . . . . . . . . . . . . . . . . . . 259
Fig. 13.3 Geometry of a single scattering. The components Ir
and Il are respectively perpendicular and parallel to the
scattering plane (, ) . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 262
List of Figures xxv
xxix
xxx List of Tables
Table 18.1 The matrices G̃(z), H(z), G̃R (z), and HR (z) at the origin
and at infinity. The matrix R = GR (∞) is a rotation
matrix, which can be calculated by solving numerically
the nonlinear integral equations for G̃R (z) or HR (z). The
matrix E = diag[, Q ] is equal to L(0). The I-matrix is
defined by I(0, z) = HR (z) . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 439
Table 19.1 The asymptotic behaviors of K(τ ) for τ → ∞, g(ξ ) for
ξ → ∞, k(ν) for ν → 0, and K̂(k) for k → 0. The
kernel is an even function of τ . Here τ , ξ , and ν are
positive. For monochromatic scattering, k(ν) is zero in
the interval [0, 1[ . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 465
Table 19.2 The index α for the Doppler and Voigt profiles and the
asymptotic expressions of f (y) and f (y) for y large.
The function f (y) is the inverse function of the profile
ϕ(x) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 466
Table 22.1 Infinite medium. Behavior at large optical depths of the
resolvent function for small values of , the destruction
probability per scattering. The constant a is the Voigt
parameter of the line. The nearly conservative zone
corresponds to 1 τ < τeff and the strong absorption
zone to τ > τeff . For monochromatic scattering τeff ∼ ν0 . . . . . . . . 509
Table 22.2 Semi-infinite medium. Asymptotic behavior of the
resolvent function in the nearly conservative and strong
absorption zones. The thermalization lengths τeff are
given in Table 22.1. For monochromatic scattering
τeff ∼ ν0 . For = 0, the monochromatic √ resolvent
function has exactly the value 3 at infinity . .. . . . . . . . . . . . . . . . . . . . 513
Table 24.1 Moments of Hl (μ) and Hr (μ) and other numerical
constants. The two first moments of Hl (μ) and Hr (μ),
the constants q and c, and the values of Hl (1) and Hr (1)
are from Chandrasekhar (1960, p. 248). This reference
provides also tables for Hl (μ) and Hr (μ), μ ∈ [0, 1],
used to calculate numerically the third and fourth order
moments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 544
Table 25.1 Scaling laws for complete frequency redistribution
with a Doppler profile. In the first column, the small
expansion parameter; in row 1: β = 0 and T = ∞, in
row 2: = 0 and T = ∞, and in row 3: β = = 0. The
four subsequent columns show τeff , the characteristic
scale of variation of the radiation field, xc , the
characteristic frequency defined by τeff ϕ(xc ) ∼ 1, the
mean number of scatterings N, and the mean path
length L . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 559
xxxii List of Tables