Professional Documents
Culture Documents
Full Chapter Resonant Tunneling Quantum Waveguides of Variable Cross Section Asymptotics Numerics and Applications 2Nd Edition Lev Baskin PDF
Full Chapter Resonant Tunneling Quantum Waveguides of Variable Cross Section Asymptotics Numerics and Applications 2Nd Edition Lev Baskin PDF
https://textbookfull.com/product/a-cross-section-of-
psychological-research-journal-articles-for-discussion-and-
evaluation-2nd-edition-andrea-k-milinki/
https://textbookfull.com/product/elasticity-theory-applications-
and-numerics-fourth-edition-sadd/
https://textbookfull.com/product/finite-mathematics-as-the-
foundation-of-classical-mathematics-and-quantum-theory-with-
applications-to-gravity-and-particle-theory-1st-edition-felix-
lev/
https://textbookfull.com/product/elasticity-theory-applications-
and-numerics-4th-edition-martin-h-sadd/
Temporal Quantum Correlations and Hidden Variable
Models 1st Edition Costantino Budroni (Auth.)
https://textbookfull.com/product/temporal-quantum-correlations-
and-hidden-variable-models-1st-edition-costantino-budroni-auth/
https://textbookfull.com/product/geoengineering-the-anthropocene-
and-the-end-of-nature-jeremy-baskin/
https://textbookfull.com/product/how-to-love-yourself-the-art-of-
loving-myself-2nd-edition-jonathan-kis-lev/
https://textbookfull.com/product/health-promotion-2nd-edition-
cross/
https://textbookfull.com/product/sustainable-tunneling-and-
underground-use-mona-badr/
Lecture Notes on Numerical Methods
in Engineering and Sciences
Lev Baskin
Pekka Neittaanmäki
Boris Plamenevskii
Oleg Sarafanov
Resonant Tunneling
Quantum Waveguides of Variable
Cross-Section, Asymptotics, Numerics,
and Applications
Second Edition
Lecture Notes on Numerical Methods
in Engineering and Sciences
Editorial Board
Francisco Chinesta, Ecole Centrale de Nantes, Nantes Cedex 3, France
Series Editor
Eugenio Oñate, Jordi Girona, 1, Edifici C1 - UPC, Universitat Politecnica de
Catalunya, Barcelona, Spain
Editorial Board
Charbel Farhat, Dept. Mechanical Engineering, Stanford University, Stanford, CA,
USA
C.A. Felippa, Department of Aerospace Engineering Scie, Univ of Colorado, Coll
of Engg & App, Boulder, CO, USA
Antonio Huerta, Universitat Politècnica de Cataluny, Barcelona, Spain
Thomas J. R. Hughes, Institute for Computational Enginee, University of Texas at
Austin, Austin, TX, USA
Sergio Idelsohn, CIMNE - UPC, Barcelona, Spain
Pierre Ladevèze, Ecole Normale Supérieure de Cachan, Cachan Cedex, France
Wing Kam Liu, Evanston, IL, USA
Xavier Oliver, Campus Nord UPC, International Center of Numerical M,
Barcelona, Spain
Manolis Papadrakakis, National Technical University of At, Athens, Greece
Jacques Périaux, Barcelona, Spain
Bernhard Schrefler, Mechanical Sciences, CISM - International Centre for, Padua,
Italy
Genki Yagawa, School of Engineering, University of Tokyo, Tokyo, Japan
Mingwu Yuan, Beijing, China
This series publishes text books on topics of general interest in the field of
computational engineering sciences.
The books will focus on subjects in which numerical methods play a fundamental
role for solving problems in engineering and applied sciences. Advances in finite
element, finite volume, finite differences, discrete and particle methods and their
applications to classical single discipline fields and new multidisciplinary domains
are examples of the topics covered by the series.
The main intended audience is the first year graduate student. Some books define
the current state of a field to a highly specialised readership; others are accessible to
final year undergraduates, but essentially the emphasis is on accessibility and
clarity.
The books will be also useful for practising engineers and scientists interested in
state of the art information on the theory and application of numerical methods.
Resonant Tunneling
Quantum Waveguides of Variable
Cross-Section, Asymptotics, Numerics,
and Applications
Second Edition
123
Lev Baskin Pekka Neittaanmäki
Department of Mathematics Mathematical Information Technology
St. Petersburg State University University of Jyväskylä
of Telecommunications Jyväskylä, Finland
St. Petersburg, Russia
Oleg Sarafanov
Boris Plamenevskii Department of Higher Mathematics
Department of Higher Mathematics and Mathematical Physics
and Mathematical Physics St. Petersburg State University
St. Petersburg State University St. Petersburg, Russia
St. Petersburg, Russia
This Springer imprint is published by the registered company Springer Nature Switzerland AG
The registered company address is: Gewerbestrasse 11, 6330 Cham, Switzerland
Preface
Devices based on the phenomenon of electron resonant tunneling are widely used in
electronics. Efforts are directed toward refining properties of resonance structures.
There are prospects for building new nanosize electronic elements based on
quantum dot systems. However, the role of resonance structure can also be given to
a quantum wire of variable cross-section. Instead of an “electrode—quantum dot—
electrode” system, one can use a quantum wire with two narrows. A waveguide
narrow is an effective potential barrier for longitudinal electron motion along a
waveguide. The part of the waveguide between two narrows becomes a “resonator,”
where electron resonant tunneling can occur. This phenomenon consists in the fact
that, for an electron with energy E, the probability TðEÞ to pass from one part of the
waveguide to the other part through the resonator has a sharp peak at E ¼ Eres ,
where E res denotes a “resonant” energy. Such quantum resonators can find appli-
cations as elements of nanoelectronic devices and provide some advantages in
regard to operation properties and production technology.
In the book, we study electron resonant tunneling in two- and three-dimensional
quantum waveguides of variable cross-sections in the time-independent approach.
We suggest mathematical models for the resonant tunneling and develop asymp-
totic and numerical approaches for investigating the models. We also present
schemes for several electronic devices based on the phenomenon of resonant tun-
neling. The book is addressed to mathematicians, physicists, and engineers inter-
ested in waveguide theory and its applications in electronics.
v
vi Preface
The second edition of the book contains new subjects: the resonant tunneling in
2D waveguides with several resonators (Chap. 6) and the resonant tunneling of
high-energy electrons in 2D waveguides (Chap. 7). The discussion of a finite
waveguide work function given in the first edition is supplemented by a study of the
impact of a finite work function on transport of high-energy electrons and is now
presented in Chap. 9. Several mistakes (see Rem. 5.5.2) and a number of misprints
are corrected.
1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
2 Waveguides. Radiation Principle. Scattering Matrices . . . . . . . . . . 17
2.1 Boundary Value Problem in a Cylinder . . . . . . . . . . . . . . . . . . 19
2.1.1 Statement of the Problem. Operator Pencil . . . . . . . . . . 19
2.1.2 The Solvability of the Problem in a Cylinder . . . . . . . . 24
2.1.3 Asymptotics of Solutions . . . . . . . . . . . . . . . . . . . . . . 25
2.2 Problem in a Domain G with Cylindrical Ends . . . . . . . . . . . . . 29
2.2.1 Statement and Fredholm Property of the Problem . . . . . 29
2.2.2 Asymptotics of Solutions . . . . . . . . . . . . . . . . . . . . . . 30
2.2.3 Properties of the Index Ind Ab ðlÞ and of the Spaces
kerAb ðlÞ and cokerAb ðlÞ . . . . . . . . . . . . . . . . . . . . . 31
2.2.4 Calculation of the Coefficients in the Asymptotics . . . . 33
2.3 Waves and Scattering Matrices . . . . . . . . . . . . . . . . . . . . . . . . 36
2.3.1 Waves . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36
2.3.2 Continuous Spectrum Eigenfunctions. The Scattering
Matrix . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 38
2.3.3 The Intrinsic Radiation Principle . . . . . . . . . . . . . . . .. 40
3 Properties of Scattering Matrices in a Vicinity of Thresholds . . . . . 43
3.1 Augmented Space of Waves . . . . . . . . . . . . . . . . . . . . . . . . . . 44
3.1.1 Waves in a Cylinder . . . . . . . . . . . . . . . . . . . . . . . . . . 44
3.1.2 Waves in Domain G . . . . . . . . . . . . . . . . . . . . . . . . . . 47
3.2 Continuous Spectrum Eigenfunctions. Scattering Matrices . . . . . 48
3.2.1 Intrinsic and Expanded Radiation Principles . . . . . . . . . 48
3.2.2 Analyticity of Scattering Matrices with Respect
to Spectral Parameter . . . . . . . . . . . . . . . . . . . . . .... 53
3.3 Other Properties of the Scattering Matrices . . . . . . . . . . . . .... 58
3.3.1 The Connection Between SðlÞ and SðlÞ
for ¿ 0 \l\¿ . . . . . . . . . . . . . . . . . . . . . . . . . . . .... 58
vii
viii Contents
Resonance Structures
2
− (x) + U (x)(x) = E(x), −∞ < x < +∞, (1.0.1)
2m
where U (x) = U1 for x ∈ [x1 , x2 ], U (x) = U2 for [x3 , x4 ], U1 and U2 are positive
constants, and x1 < x2 < x3 < x4 ; moreover, U (x) = 0 for the rest x. The parts of
U over [x1 , x2 ] and [x3 , x4 ] are called potential barriers and [x2 , x3 ] is a potential
well; the barriers and the well comprise a resonator. An electron wave function
satisfies equation (1.0.1), where E is the electron energy, m is the electron mass, and
U is the electron energy. Besides, can be chosen to satisfy the equalities (see, e.g.,
[3, 15, 20]) ikx
e + r e−ikx as x < x1 ,
(x) =
teikx as x > x4 ,
where k = (2m E/2 )1/2 . For x < x1 , functions eikx and r e−ikx are considered as an
incoming wave and a reflected wave, respectively, and, for x > x4 , teikx is a transmit-
ted wave. The values T (E) = |t (E)|2 and R(E) = |r (E)|2 are called a transmission
coefficient and a reflection coefficient. It turns out that T (E) + R(E) = 1; the T (E)
(R(E)) is interpreted as a probability for the electron to transmit through the res-
onator (to be reflected from the resonator). Under certain conditions, there exists a
“resonance” Er es , 0 < Er es < min{U1 , U2 }, such that for E = Er es the transmission
coefficient T takes a maximal value (in particular, T (Er es ) = 1 can be the case).
Then with probability close to 1 the electron transmission can take place, under the
barriers, through the resonator. This phenomenon is called resonant tunneling.
ε
1
Scattering Matrix
functions satisfy (1.0.2); moreover, the functions are bounded and do not vanish at
infinity. To describe the wave function behavior at infinity, we will use solutions to
the problem in the cylinder
n
y,t = y + ∂t2 , y = ∂ 2j , ∂ j = ∂/∂ y j .
j=1
− ( y + τ )ϕ(y) = 0, y ∈ , (1.0.4)
ϕ(y) = 0, y ∈ ∂,
and set
u± ∓ −1/2
1 (y, t) = (2|λ1 |) exp(iλ∓
1 t)ϕ1 (y) (1.0.5)
with λ±
1 = ±(μ − τ1 )
1/2
. Functions (1.0.5) are bounded, satisfy (1.0.3), and do not
decay at infinity. We will call the u + −
1 (u 1 ) a wave incoming from +∞ (outgoing to
+∞).
For μ ∈ (τ2 , τ3 ), besides u ±
1 in (1.0.5), we have waves of the form
u ± (y, t) = (2|λ∓
2 |)
−1/2
exp(iλ∓
2 t)ψ(y), (1.0.6)
where λ± 2 = ±(μ − τ2 )
1/2
and ψ is an eigenfunction of problem (1.0.4) correspond-
ing to τ2 . The number of pairs of the form (1.0.6) is equal to the multiplicity κ(τ2 )
of the eigenvalue τ2 ; as an eigenfunction ψ, the elements ψ1 , . . . , ψκ(τ2 ) of a basis
in the eigenspace of problem (1.0.4) have to be chosen, subject to the orthogonality
and normalization conditions
ψ p (y)ψq (y) dy = δ p,q , p, q = 1, . . . , κ(τ2 ).
In general, for μ ∈ (τl , τl+1 ), the number of the wave pairs in the cylinder is equal
to κ(τ1 ) + · · · + κ(τl ).
Let μ be different from the eigenvalues of problems (1.0.4) in 1 , . . . , T . Given
μ, we enumerate all wave pairs in the cylinders 1 , . . . , T by the same index
j = 1, 2, . . . , M. Among electron wave functions in G, there exist 1 , . . . , M that
admit the representations
M
l (x) = u l+ (x) + Sl j u −j (x) + O(exp (−δ|x|)) (1.0.7)
j=1
Remark 1.0.1 For the one-dimensional resonance structure (1.0.1) there are two
related scattering channels: u + − + − +
1 , u 1 and u 2 , u 2 , where u 1 (x) = e
ikx
(u −
1 (x) = e
−ikx
)
+
is an incoming (outgoing) wave to the left of the resonator, and u 2 (x) = e−ikx
(u −
2 (x) = e
ikx
) is an incoming (outgoing) wave to the right of the resonator. Thus,
the scattering matrix is of size 2 × 2.
Next we are going to state the method employed for numerical simulation of resonant
tunneling. In the Introduction, we restrict ourselves to considering the scattering
matrix on a finite interval of the continuous spectrum containing no thresholds. In
Chap. 4, we modify the method to calculate the scattering matrix also in vicinity of
thresholds and present a justification for the method in both of these situations.
Introduce the notation
r,R r r R N r,R
+ = {(y , t ) ∈ : t > R}, G = G \ ∪r =1 +
r r
− ( + μ)XlR = 0, x ∈ G R ;
XlR = 0, x ∈ ∂G R \ R ;
M
(∂ν + iζ)XlR = (∂ν + iζ)(u l+ + a j u −j ), x ∈ R , (1.0.9)
j=1
M
(∂ν + iζ)l = (∂ν + iζ)(u l+ + a j u −j ) + O(e−γ R )
j=1
M
JlR (a1 , . . . , a M ) = XlR − u l+ − a j u −j ; L 2 ( R )2 , (1.0.10)
j=1
where XlR is a solution to problem (1.0.9). One can expect that a 0j (R, μ) → Sl j (μ)
at exponential rate as R → ∞ and j = 1, . . . , M.
Let [μ , μ ] be an interval of the continuous spectrum without thresholds. In
Chap. 4, we prove, in particular, that for all R ≥ R0 and μ ∈ [μ , μ ] there exists
a unique minimizer a(R, μ) = (a1 (R, μ), . . . , a M (R, μ)) of functional (1.0.10) and
the estimates
O
ω
We denote by K a double cone with vertex at the origin O that contains the x-axis
and is symmetric about the coordinate axes. The set K ∩ S 1 , where S 1 is a unit circle,
consists of two simple arcs. Assume that contains the cone K and a neighborhood
of its vertex; moreover, outside a large disk (centered at the origin) coincides with
K . The boundary ∂ of is supposed to be smooth (see Fig. 1.3).
Denote by (ε) the domain obtained from by the contraction with center at O
and coefficient ε. In other words, (x, y) ∈ (ε) if and only if (x/ε, y/ε) ∈ . Let
K j and j (ε) stand for K and (ε) shifted by the vector r j = (x 0j , 0), j = 1, 2. We
assume that |x10 − x20 | is sufficiently large so the distance from ∂ K 1 ∩ ∂ K 2 to G is
positive. We put G(ε) = G ∩ 1 (ε) ∩ 2 (ε) (Fig. 1.4).
The wave function of a free electron of energy k 2 satisfies the boundary value
problem
The problems
where j = 0, 1, 2, are called the first kind limit problems. In the domains j , j =
1, 2, we consider the boundary value problems
w(ξ j , η j )= F(ξ j , η j ), (ξ j , η j ) ∈ j ,
w(ξ j , η j )= 0, (ξ j , η j ) ∈ ∂ j ,
which are called the second kind limit problems; (ξ j , η j ) are Cartesian coordinates
with origin at O j .
We denote by ke2 a simple eigenvalue of problem (1.0.12) in the resonator G 0
and by kr2 (ε) a resonance frequency such that kr2 (ε) → ke2 as ε → 0. For |k 2 − kr2 | =
O(ε2π/ω ) the asymptotic representations hold:
1
T (k, ε) = 1 + O(ε2−τ ) ,
k − kr2 2
2
1 + P2
ε4π/ω
kr2 (ε) = ke2 + Qε2π/ω + O ε4π/ω ,
1
ϒ(ε) = ε4π/ω 1 + O(ε1−τ /2 ) ,
P
where T (k, ε) is the electron transmission coefficient and ϒ(ε) is the width of the
resonant peak at its half-height (which is inversely proportional to the resonator
quality factor), τ being an arbitrarily small positive number; the P and Q are the
products of several constants in the asymptotics of limit problem solutions near
corners or at infinity.
Without numerical values of the constants, the asymptotic formulas provide only a
qualitative picture. To find the constants, one has to solve numerically several bound-
ary value problems. We state the problems and describe a way to solve them. When
the constants are found, the asymptotics can be used as an approximate solution.
10 1 Introduction
However, it remains uncertain for what band of parameters the approximation is reli-
able. On the other hand, one should expect numerical approach to be efficient only if
the waveguide narrows are not too small in diameter and if the resonant peak of the
transmission coefficient is sufficiently wide. Therefore a detailed picture of resonant
tunneling can be achieved when the asymptotic and numerical approaches are com-
bined. Independently of asymptotic approach, an approximation to the waveguide
scattering matrix is calculated. For that purpose, we employ the method from Chap.
4. Then we can compare the asymptotics with calculated constants and the scattering
matrix (the transition and reflection coefficients). It turns out, that there is an interval
for ε, where the asymptotic and numerical results practically coincide. To the right
of the interval, the asymptotics vanishes but the numerical method for calculation of
the scattering matrix is effective; to the left of the interval, the numerical method is
ill-conditioned while the asymptotics is reliable.
2
kr,l (ε) = ke2 − Q l ε2π/ω + O(ε4π/ω ),
πl
Q l = f + 2g cos .
n+1
1
T (k, ε) = 1 + O(ε4π/3ω ) ,
n 2 2
l=1 k − kr,l
2
1+ P2
ε2(n+1)π/ω
4 πl 4π/ω
ϒl = sin2 ε + O(ε6π/ω ).
(n + 1)g n−1 P n+1
Pε8π/ω (k 2 − ke2 )2
T (k, ε) = + O(ε4π/ω ),
[(k 2 − 2 2
kr,1 ) + Q 1 ε8π/ω ] [(k 2 − kr,2
2 2
) + Q 2 ε8π/ω ]
the constants ω, P, Q j , and f j are positive, they depend only on the waveguide
geometry; τ is a positive number. Note that the leading term of the asymptotics
vanishes at k = ke , i.e. the resonant reflection occurs, which was not observed in the
case of a simple eigenvalue of the resonator.
to state k. We denote by Tnk (E) the transmission coefficient of the wave, calculate
the dependence E → Tnk (E) by computing the scattering matrix S(E), and obtain
Tnk (E) = |Snk (E)|2 , where Snk (E) is the entry of S(E). The curve E → Tnk (E)
can be sufficiently complicated and not always easily interpreted. To explain the
curve, we consider Snk (E) as a probability amplitude and represent it in the form
Snk (E) = s Ansk (E), where Ansk (E) is the probability amplitude of the transmis-
sion from n to k through an intermediate state s; the summation is over all intermediate
states (cf. [21]).
As before, we denote by G(ε) a waveguide with two narrows and let G 0 be
the closed resonator, that is, the bounded part of the limit waveguide G(0)(see
Fig. 1.5); generally, the resonator form may be arbitrary. We denote by k12 ≤ k22 ≤ . . .
the eigenvalues of problem (1.0.12) with j = 0 numbered according to their mul-
tiplicities. Then the resonant energies of the waveguide G(ε) form the sequence
ReE 1 , ReE 2 , . . . , where E 1 , E 2 , . . . can be viewed as the “perturbed” k12 , k22 , . . .
and ImE j < 0 for all j = 1, 2, . . . . The amplitude Ansk admits the representation
(s)
(s) Rnk (E)
Ansk (E) = Hnk (E) +
E − Es
(s) (s)
with continuous functions E → Hnk (E) and E → Rnk (E). In a small neighborhood
of ReEr ,
Rnk (Er ) 2
|Snk (E)|2 = | Ansk (E)|2 ≈ |Hnk (Er ) + | ≡ Tnk (E),
s
E − Er
where Hnk (Er ) and Rnk (Er ) are constant. We take the function Tnk (E) as an approx-
imation to the calculated |Snk (E)|2 and find the constants Hnk (Er ), Rnk (Er ), and Er
by the method of least squares.
2 −1
1 1 2 k2 − k2
T1 (k, ε1 , ε2 ) = z+ + P 2 2μ11 +1 2μr21 +1 1 + O(ετ11 + ετ22 ) ,
4 z ε1 ε2
2μ +1 2μ +1
where τ j are the same as in (10.1.6), z = Qε1 11 /ε2 21 , while P and Q are
constant. The width of the resonant peak at its half-height (calculated for the principal
part in the asymptotics of T ) is
1 1 2μ11 +1 2μ21 +1
ϒ(ε1 , ε2 ) = z+ ε ε2 (1 + O(ετ11 + ετ22 )).
P z 1
The presence of a magnetic field can essentially affect the basic characteristics of
the resonant tunneling and bring new possibilities for applications in electronics. In
particular, in the presence of a magnetic field, the tunneling phenomenon is feasible
for producing spin-polarized electron flows consisting of electrons with spins of the
same direction. In Chaps. 11 and 12 we consider the same 2D and 3D waveguides with
narrows as in Chaps. 5 and 10, respectively. A part of the resonator is occupied by a
homogeneous magnetic field. An electron wave function satisfies the Pauli equation
in a waveguide and vanishes on its boundary. An electron energy is in between
the first and the second thresholds. The asymptotics of basic resonant tunneling
characteristics are presented as the narrow diameters tend to zero. Moreover, in
Chap. 11, the asymptotic results for 2D waveguides are compared with numerical
ones obtained by approximate computing the scattering matrix; there is an interval
of ε (the narrow diameter) where the asymptotic and numerical results practically
coincide. Using the approximate scattering matrix, we also observe the dependence
of the tunneling characteristics on a magnetic field position in the resonator.
1 Introduction 15
In Chap. 13, for electrons of high energy, we generalize the asymptotic theory exposed
in Chap. 10. We present and justify the asymptotics of tunneling characteristics as
the narrow diameters tend to zero.
First, we briefly outline the chapter content. Section 2.1 is devoted to the boundary
value problem
u (y, λ) =
(− y + λ2 − μ) f (y, λ), y ∈ ,
u (y, λ) = 0, y ∈ ∂. (2.0.3)
If the inverse operator A(λ, μ)−1 of problem (2.0.3) exists for all λ ∈ R, the μ being
fixed, we obtain a solution u to problem (2.0.1) of the form
+∞
u(·, t) = (2π) −1/2
exp (iλt)A(λ, μ)−1
f (·, λ) dλ. (2.0.4)
−∞
However, the spectrum of the pencil λ → A(λ, μ), that is, the set of numbers λ such
that the operator A(λ, μ) is not invertible, consists of an imaginary number sequence
accumulating at infinity and, for sufficiently large μ, additionally contains finitely
many real numbers. Therefore, formula (2.0.4) can fail and we will use the complex
Fourier transform
Let us assume that the line R + iβ is free from the spectrum of A(·, μ) and the f in
(2.0.1) satisfies the condition
exp (2βt)| f (y, t)| dydt =
2
|
f (y, λ)|2 dydλ < ∞.
R+iβ
Then, according to Theorem 2.1.4, there exists a unique solution u to problem (2.0.1)
such that
u(·, t) = (2π)−1/2
exp (iλt)A(λ, μ)−1 f (·, λ) dλ (2.0.5)
R+iβ
in function spaces with weighted norms (see Fig. 1.2 and the definition of G just after
the figure). For integer l ≥ 0, we denote by H l (G) the Sobolev space with norm
l
1/2
v; H (G) =
l
|Dxα v(x)|2 d x .
j=0 G |α|= j
The “later dispatch” promised by the British cabinet was never sent.
The answer to it of the Transvaal government was, therefore,
delayed for several days, awaiting the new proposals that were to
come as the result of further deliberations on the part of Her
Majesty’s government. At last, on the eve of the outbreak of war, Mr.
Chamberlain gave out, on the 10th of October, the text of the
republic’s rejoinder to the British cabinet’s note of the 25th of
September. It was transmitted by cable, through Sir Alfred Milner,
and read thus:
“Dear Sir: The government of the South African Republic feels itself
compelled to refer the government of Her Majesty, Queen of Great
Britain and Ireland, once more to the convention of London, 1884,
concluded between this republic and the United Kingdom, and
which, in Article [242]XIV., secures certain specific rights to the white
population of this republic—namely: That all persons other than
natives, on conforming themselves to the laws of the South African
Republic—
“A—Will have full liberty, with their families, to enter, travel or reside
in any part of the South African Republic.
“This government wishes further to observe that these are the only
rights which Her Majesty’s government has reserved in the above
convention with regard to the outlander population of this republic,
and that a violation only of those rights could give that government a
right to diplomatic representations or intervention; while, moreover,
the regulation of all other questions [243]affecting the position of the
rights of the outlander population under the above-mentioned
convention is handed over to the government and representatives of
the people of the South African Republic.
“This government can only see in the above intimation from Her
Majesty’s government a new violation of the convention of London,
1884, which does not reserve to Her Majesty’s government the right
to a unilateral settlement of a question which is exclusively a
domestic one for this government, and which has been already
regulated by this government.
On the morning of August the 8th, the day after Mr. Chamberlain’s
warlike speech, the London papers announced that the Liverpool
and Manchester regiments, then at the Cape, had been ordered to
Natal; that the Fifteenth Hussars were to embark on the 23d of
August, and that troops were to be massed along the Transvaal
frontier. On the 11th of August it was announced that 12,000 British
troops were to be dispatched from India to South Africa, and on the
same day a large consignment of war stores, including medical
requisites, was given out from the royal arsenal, Woolwich, for
shipment to Natal, and the sum of $2,000,000 in gold was sent to
South Africa for the War Office account. British troops began to
arrive in South Africa from India and from England in the first week of
October. By the 10th some 15,000 had landed. [249]These were
hurried to the frontiers of the Orange Free State—both west and east
—most of them being concentrated along the northern boundary of
Natal, convenient to the southern frontier of the Transvaal.
Perceiving that all the days spent in waiting for that “later dispatch”
were being used by Great Britain in massing her gigantic powers of
war in South Africa and along the Transvaal frontier, and believing
that no such dispatch would now come until the points of war were
all [251]secured by his great antagonist, President Kruger at last
astonished the world—and, most of all, Great Britain—by issuing an
ultimatum sufficiently bold and defiant to have come from any of the
first-rate powers of the earth.
“Fourth—That Her Majesty’s troops which are now on the high seas
shall not be landed in any part of South Africa.
Within an hour the telegraphic wires had flashed through all the
South African Republic the ominous word “Oorlog”—war!
Mr. Conyngham Greene at once asked for his passport, and on the
next day, October the 12th, with his family, he was sent, attended by
a guard of honor, to the border of the Orange Free State, [254]where
a similar guard received and conducted him to British territory in
Cape Colony.
Thursday, the 12th of October, was a busy and exciting day in both
the Transvaal and the Orange Free State—for the two republics
stood as one in the struggle. That night—twenty-four hours after war
had been declared—30,000 burghers were on the borders ready to
do battle. Of these 20,000 invaded Natal under General Joubert, and
the vanguard under General Kock occupied Newcastle on the 13th
of October. The other 10,000, under General Piet Cronje, crossed
the western border into British Bechuanaland and marched on
Mafeking.
Thus, and for the causes set forth, began the Africanders’ Second
War of Independence. It was not in the proposed scope of this book
to treat of its fortunes. The prospect is that it will be a long and
sanguinary war. The story of it will afford abundant and interesting
matter for a later volume.
It only remains to show that in all the matters in dispute between the
government of the Transvaal and that of Great Britain, and in the war
which resulted therefrom, the two Africander republics acted in
solidarity. Early in November, 1899, the President of the Orange
Free State [255]announced this to his people and to the world in the
following proclamation:
“With our sister republic we are not only bound by ties of blood, of
sympathy and of common interests, but also by formal treaty which
has been necessitated by circumstances. This treaty demands of us
that we assist her if she should be unjustly attacked, which we
unfortunately for a long time have had too much reason to expect.
We therefore cannot passively look on while injustice is done her,
and while also our own dearly bought freedom is endangered, but
are called as men to resist, trusting the Almighty, firmly believing that
He will never permit injustice and unrighteousness to triumph.
“Now that we thus resist a powerful enemy, with whom it has always
been our highest desire [256]to live in friendship, notwithstanding
injustice and wrong done by him to us in the past, we solemnly
declare in the presence of the Almighty God that we are compelled
thereto by the injustice done to our kinsmen and by the
consciousness that the end of their independence will make our
existence as an independent state of no significance, and that their
fate, should they be obliged to bend under an overwhelming power,
will also soon after be our own fate.
“Those plans have now reached their climax in the open violence to
which the present British government now resorts. While we readily
acknowledge the honorable character of thousands of Englishmen
who loathe such deeds of robbery and wrong, we cannot but abhor
the shameless breaking of treaties, the feigned pretexts for the
transgression of law, the violation of international law and of justice
and the numerous right-rending deeds of the British statesmen, who
will now force a war upon the South African Republic. On their heads
be the guilt of blood, and may a just Providence reward all as they
deserve.
“Burghers of the Orange Free State, rise as one man against the
oppressor and the violator of right!
“In the strife to which we are now driven have care to commit no
deed unworthy of a Christian and of a burgher of the Orange Free
State. Let us look forward with confidence to a fortunate end of this
conflict, trusting to the [260]Higher Power without whose help human
weapons are of no avail.
“May He bless our arms. Under His banner we advance to battle for
liberty and for fatherland.
Nature has divided that part of Africa lying south of the Zambesi River into three
distinct and well-defined regions. A strip of lowland skirts the coast of the Indian
Ocean all the way from Cape Town around to Natal, Delagoa Bay and still
northeast to the mouths of the Zambesi. Between Durban, the principal port of
Natal, and Cape Town this strip is very narrow in places—the hills coming down
almost to the margin of the sea. Beyond Durban, to the northeast, the low plain
grows wider. This belt of lowland is more or less swampy, and from Durban
northward is exceedingly malarious and unhealthful. This feature is a prime
factor in the physical [262]structure of the country and has had much to do with
shaping its history.
The second region is composed of the elevated and much broken surface
presented by the Drakensburg or Quathlamba range of mountains, reaching
from Cape Town to the Zambesi Valley—a distance of sixteen hundred miles. In
traveling inland, after leaving the level belt, at from thirty to sixty miles from the
sea the hills rise higher and higher—from three thousand to six thousand feet.
These hills are only the spurs of the principal range, some of whose peaks rise
to an elevation of eleven thousand feet.
Beyond the Quathlamba Mountains, to the west and north, is the third natural
division of South Africa—a vast tableland or plateau, varying from three
thousand to five thousand feet above the sea level. This region occupies about
seven-eighths of the area of South Africa.
Geologically considered, the oldest formation is found in the northern part of the
tableland and toward the northeastern end of the Quathlamba [263]Mountains.
The principal formations in this region are granite and gneiss, believed to be of
great antiquity—probably of the same age as the Laurentian formations in
America. The rocks of the Karoo district are not so ancient. There are no traces
anywhere in South Africa of late volcanic action, nor has any active volcano
been discovered there; but eruptive rocks of ancient date—porphyries and
greenstones—are found overlying the sedimentary deposits in the Karoo district
and in the mountain systems of Basutoland and the Orange Free State.
The African coast is notably poor in harbors. There is no haven between Cape
Town and Durban. From Durban to the Zambesi there are but two good ports—
that of Delagoa Bay and Beira. With the exception of Saldanha Bay, twenty
miles north of Cape Town, the western coast, for a thousand miles, has no
harbor.
The temperature in Southern Africa is much lower than the latitude would lead
one to expect. This is accounted for by the fact that there is a vast
preponderance of water in the southern hemisphere, which has the effect of
giving a cooler temperature than prevails in a corresponding northern latitude.
The difference in both heat and cold represents over two degrees of difference
[264]in latitude. Thus, Cape Town, 34° S., has a lower temperature in both
summer and winter than Gibraltar and Aleppo, in 36° N. Nevertheless, the
thermometer registers high in some parts of South Africa. Even at Durban, in
latitude 30° S., the heat is often severe, and the northern part of the Transvaal
and the British territories to the north of it lie within the Tropic of Capricorn. The
mean temperature in South Africa proper is 70° Fahrenheit in January and 80° in
July.
Over most of the country the climate is exceptionally dry. In the region of Cape
Colony there are well-defined summer and winter; but in the rest of South Africa
for about two-thirds of the year there is only a dry season, when the weather is
cooler, and a wet season of four or five months, when the sun is the highest and
the heat is most intense. The rainy season is not so continuous, nor is there so
great a precipitation, as in some other hot countries. In the parts where the
rainfall is heaviest, averaging over thirty inches in the year, the moisture soon
disappears by evaporation and absorption, and the surface remains parched till
the next wet season. As a consequence of this the air is generally dry, clear and
stimulating.
CATTLE ON THE VAAL RIVER.
[265]
It is interesting to note the effect upon climate of the physical structure described
above. The prevailing and rain-bringing winds are from the east and the
southeast. They bring sufficient moisture to the low plain along the sea coast,
and passing inland the rain-bearing clouds water the foothills of the Quathlamba
Mountains and precipitate snow on the loftier peaks beyond them. A portion of
the moisture is carried still farther to the west and falls in showers on the eastern
part of the plateau—the Transvaal, the Orange Free State, the eastern border of
Bechuanaland and the region northward toward the Zambesi. Sections farther to
the north and west receive but little of the annual rainfall, ranging from five to ten
inches in the year. That little is soon dissipated, the surface becomes dry and
hard, and such vegetation as springs up under the brief showers soon dies.
Much of this region is a desert, and so must remain until more and more
continuous moisture is supplied, either by artificial irrigation or by some
favorable change in natural conditions.
From these permanent physical features—the lowlands along the coast, the
elevated plateau in the interior, the mountain range running between them, a
burning sun and a dry atmosphere[266]—have developed many of the other
natural phenomena of South Africa.
The rivers of that country—laid down in great numbers on the maps—are not
rivers during much of the year. In the dry season they are either without water
altogether or consist of a succession of little pools scarcely sufficient to supply
the cattle on their banks with drink. And when they are rivers they are, most of
the time, such as can neither be forded nor navigated; the violent rains—
continuing for hours and sometimes for many days—have converted them into
roaring torrents.
Now, if that country could have been entered by waterways, as were North and
South America, it would not have remained an unknown land so long. But there
was no other means of penetrating it than the lumbering ox-wagon, making at
best a dozen miles a day, with frequent long halts in the neighborhood of good
grass in order to rest and recuperate the cattle. It is this lack of navigable rivers
that now compels the people to depend exclusively on railways for internal
transportation and travel. With the exception of tidal streams there is no internal
water communication of any value.
Another peculiarity of the east coast rivers [267]arises out of the nearness of the
Quathlamba Mountains to the sea. Such rivers as take their rise in the
mountains have very short courses, and the few that come from beyond, finding
channels through the mountain passes, are so obstructed by rapids and
cataracts at the point of descent from the higher levels that no boat can ascend
them.
South Africa presents to the foreigner from cooler climates no serious danger as
to health. The sun-heat would be trying were it not for the dryness of the
atmosphere and the invariable coolness of the nights, which have the effect of a
refreshing tonic. With due care in providing sufficient wraps for the occasional
cold day in the dry season, and the means of comfortable sleep during the cool
nights, there is nothing to fear.
The much-dreaded malarial fever has its habitat in the lowlands of both the east
and the west coast. Persons who are not immune to it can choose their place of
residence on the higher lands, or take refuge in quinine.
The dryness and purity of the air in many parts of South Africa—notably Ceres,
Kimberley, Beauport West and other places in the interior plateau—make it
peculiarly suitable for [268]persons suffering from any form of chest disease—
always excepting tuberculosis, for which the sure remedy has not yet been