Thai Journal of Mathematics

Special Issue (Annual Meeting in Mathematics, 2010) : 83–97

www.math.science.cmu.ac.th/thaijournal
Online ISSN 1686-0209

One Dimensional Scattering Problems:

A Pedagogical Presentation of the
Relationship between Reflection and
Transmission Amplitudes

P. Boonserm and M. Visser

Abstract : In this article we provide a pedagogical introduction to scattering

theory in one space dimension. This is an elegant topic that is mathematically
simple and physically transparent. We shall apply the Schrödinger equation to a
generic system to identify the Bogoliubov coefficients. Furthermore, we shall then
derive a number of significant relationships between reflection and transmission
amplitudes.

Keywords : Schrödinger equation, Scattering problems, Bogoliubov coefficients,

Reflection and transmission amplitudes.
2000 Mathematics Subject Classification : Mathematical Physics; Quan-
tum Physics; High-Energy Physics.

1 Introduction

In this article we shall present a simple pedagogical introduction to quantum

scattering theory in one space dimension. This is a beautiful subject that is math-
ematically simple and physically transparent. Moreover, it still leads to important
and significant novel results [1, 2, 3, 4, 5, 6, 7].
One-dimensional scattering problems appear in a vast variety of physical con-
texts, textbook presentations [8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21,
22, 23], and research monographs [24, 25, 26, 27, 28, 29, 30, 31]. For instance,
in acoustics one might be interested in the propagation of sound waves down a
long pipe, while in electromagnetism one might be interested in the physics of
wave-guides. Another important context which we want to stress in this arti-

Copyright c 2010 by the Mathematical Association of Thailand. All rights

reserved.
84 P. Boonserm and M. Visser

cle is that in quantum physics the canonical examples related to one-dimensional

scattering theory are barrier penetration and reflection. In contrast, in classi-
cal physics an equivalent problem is the analysis of parametric resonances [1].
When considering the basic ideas of “reflection and transmission amplitudes”, we
shall introduce a useful technique to derive a connection between reflection and
transmission coefficients, showing that they are related via a conceptually simple
formalism. This technique has been used multiple times in several recent related
articles [1, 2, 3, 4, 5, 6, 7].
In particular, at the end of this article we shall illustrate how to derive the
“transfer matrix” in terms of the transmission and reflection amplitudes due to
scattering by a finite-width potential well. Specifically, we are interested in the
Schrödinger equation as shown below in equation (2.1) in conditions where the
potential V (x) is zero outside of a finite interval. Purely for mathematical con-
venience we are most interested in considering potentials of compact support.
(Though much of what we will have to say will also apply to potentials with
suitably rapid falloff properties as one moves to spatial infinity.)

2 Reflection and Transmission Amplitudes

Let us consider the one-dimensional time-independent Schrödinger equation [8,

9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23]
~2 d2
− ψ(x) + V (x)ψ(x) = Eψ(x). (2.1)
2m dx2
If the potential asymptotes to a constant,
V (x → ±∞) → V±∞ , (2.2)
then in each of the two asymptotic regions there are two independent solutions to
the Schrödinger equation
exp(±ik±∞ x)
ψ± (x → ±∞) ≈ p . (2.3)
k±∞
Here the ± distinguishes right-moving modes e+ikx from left-moving modes e−ikx ,
while the ±∞ specifies which of the asymptotic regions we are in. Furthermore
r
2m (E − V±∞ )
k±∞ = . (2.4)
~
To even begin to set up a scattering problem the minimum requirements are that
potential asymptote to some constant, and this assumption will be made hence-
forth. The so-called Jost solutions (see for example [26]) are exact solutions J± (x)
of the Schrödinger equation that satisfy
exp(+ik−∞ x)
J+ (x → −∞) → p , (2.5)
k−∞
One dimensional scattering problems. . . 85
exp(+ik+∞ x) exp(−ik+∞ x)
J+ (x → +∞) → α+ p + β+ p , (2.6)
k+∞ k+∞
and
exp (−ik+∞ x)
J− (x → +∞) → p , (2.7)
k+∞
exp(−ik−∞ x) exp(+ik−∞ x)
J− (x → −∞) → α− p + β− p . (2.8)
k−∞ k−∞
There are unfortunately at least four distinct sets of conventions and formulations
p
in common use, depending on whether or not one absorbs factors of k±∞ into
the reflection and transmission amplitudes r and t respectively, and on whether
one chooses to focus on left-moving or right-moving waves as being primary. We
shall discuss three of these formulations in some detail.

2.1 Formulation 1
Letpus, for the current section, adopt the convention of not absorbing the factors
of k±∞ into r and t. We start by introducing a minor variant of Messiah’s
notation [20]
J+ (x → −∞) → t+ exp(+ik−∞ x), (2.9)
J+ (x → +∞) → exp(+ik+∞ x) + r+ exp(−ik+∞ x), (2.10)
By comparing these two different forms for the asymptotic form of the Jost function
we see that in this situation the ratios of the amplitudes are given by
1 α+ β+
p :p :p = t+ : 1 : r+ . (2.11)
k−∞ k+∞ k+∞
Thus we obtain p
β+ k+∞ β+
r+ = p = . (2.12)
k+∞ α+ α+
We also derive (in this set of conventions)
p s
1 k+∞ k+∞ 1
t+ = p = . (2.13)
k−∞ α+ k−∞ α+

Thus we have demonstrated that α+ and β+ , the (right-moving) Bogoliubov co-

efficients, are related to the (left-moving) reflection and transmission amplitudes
by s
β+ k+∞ 1
r+ = ; t+ = . (2.14)
α+ k−∞ α+
Without further calculation we can also deduce
s
β− k+∞ 1
r− = ; t− = . (2.15)
α− k−∞ α−
86 P. Boonserm and M. Visser

The explicit occurrence of k+∞ and k−∞ in these equations is an annoyance,

which is why many authors adopt the alternative normalization to be discussed
below [1, 2, 3, 4, 5, 6, 7].
In Bogoliubov language the present conventions correspond to an incoming flux
of right-moving particles (incident from the left) being amplified to amplitude α+
at a cost of a backflow of amplitude β+ . In scattering language one should consider
the complex conjugate J+∗ — this is equivalent to an incoming flux of left-moving
particles (incident from the right) of amplitude α∗+ being partially transmitted
∗
(amplitude unity) and partially scattered (amplitude β+ ). If the potential has
even parity, then the left-moving Bogoliubov coefficients are just the complex
conjugates of the right-moving coefficients, however if the potential is asymmetric
a more subtle analysis is called for.
The second interesting issue is that we can deal exclusively with α+ and β+ ,
dropping the suffix for brevity — if information about α− and β− is desired simply
work with the reflected potential V (−x). It should also be borne in mind that
the phases of β and β ∗ are physically meaningless in that they can be arbitrarily
changed simply by moving the origin of coordinates (or equivalently, physically
moving the location of the potential). The phases of α and α∗ on the other hand
do contain real and significant physical information.
For completely arbitrary potentials, with no parity restriction (so the potential
is neither even nor odd), a Wronskian analysis yields (see for example reference [20,
pages 106-108], noting that an overall minus sign between Messiah and the con-
ventions above neatly cancels):

k−∞ [1 − |r+ |2 ] = k+∞ |t+ |2 ; (2.16)

k−∞ |t− |2 = k+∞ [1 − |r− |2 ]; (2.17)

k−∞ t− = k+∞ t+ ; (2.18)

k−∞ r+ t∗+ = −k+∞ r− t∗− ; (2.19)

with equivalent relations for α and β. Then

k+∞ k−∞
T+ = |t+ |2 = |t− |2 = T− (2.20)
k−∞ k+∞

and so the barrier transmission probability is independent of direction. We also

have
phase (t+ ) = phase (t− ), (2.21)

and
phase (r+ /t+ ) = π − phase (r− /t− ), (2.22)

with equivalent relations for α and β.

One dimensional scattering problems. . . 87

2.2 Formulation 2
If we now adopt the (to our minds) more useful convention, by absorbing suit-
able factors of k+∞ and k−∞ into the definitions of r and t, then things simplify
considerably. We restart the calculation by now defining a slightly different set of
reflection and transmission amplitudes r and t via the equations
exp(+ik−∞ x)
J+ (x → −∞) → t+ p , (2.23)
k−∞
exp(+ik+∞ x) exp(−ik+∞ x)
J+ (x → +∞) → p + r+ p , (2.24)
k+∞ k+∞
By comparing these two different forms for the asymptotic form of the Jost function
we see that in this situation the ratios of the amplitudes are given by the much
simpler formulae
1 : α+ : β+ = t+ : 1 : r+ . (2.25)
We now have
β+
r+ = , (2.26)
α+
and
1
t+ = . (2.27)
α+
p
We see that by putting the factors of k±∞ into the asymptotic form of the Jost
functions, where they really belong, the formulae for r and t are suitably simplified.
For completely arbitrary potentials, with no parity restriction (so the potential
is neither even nor odd), a modified Wronskian analysis now yields (in analogy
with that reported by Messiah [20, pages 106-108]):
|t+ |2 = 1 − |r+ |2 ; (2.28)
|t− |2 = 1 − |r− |2 ; (2.29)
t− = t+ ; (2.30)
r+ t+ = −r− t∗− ;
∗
(2.31)
with equivalent relations for α and β. Then
T+ = |t+ |2 = |t− |2 = T− (2.32)
and so the barrier transmission probability is independent of direction. Because
they are independent of any overall scaling by a real number, we also retain the
previous results
phase (t+ ) = phase (t− ), (2.33)
and
phase (r+ /t+ ) = π − phase (r− /t− ), (2.34)
with equivalent relations for α and β. It is this modified set of conventions, because
they have much nicer normalization properties, that we shall prefer for the bulk
of the paper.
88 P. Boonserm and M. Visser

2.3 Formulation 3
The Schrödinger equation also can be analyzed in terms of a different formalism
based on the functions u and v, as defined by Messiah [20], and their complex
conjugates u∗ and v ∗ . Note that the Wronskian of any two such solutions is in-
dependent of x. In particular, it takes on the same value in the two asymptotic
regions. Our approach can be seen as equating these two values; we shall now de-
rive a relation between the coefficients r+ , t+ , r− , t− , or their complex conjugates.
Six such relations can be formed with the four functions u, v, u∗ and v ∗ . From
what we have seen earlier it is clear that they are very basic relations which must
be maintained whatever the form of the potential function V (x). See for instance
reference [20, pages 106–108]. Specifically, we derive (in Messiah-like conventions)

i
W (u, u∗ ) = k+∞ (1 − |r+ |2 ) = k−∞ |t+ |2 ; (2.35)
2
i
W (v, v ∗ ) = k−∞ (1 − |r− |2 ) = k+∞ |t− |2 ; (2.36)
2
i
W (u, v) = k+∞ t− = k−∞ t+ ; (2.37)
2
i
W (u, v ∗ ) = −k+∞ r+ t∗− = k−∞ r−
∗
t+ . (2.38)
2
The equations (2.35) and (2.36) are called the relations of conservation of flux.
They must always be true, and this should be verified in special cases. This name
comes from the following statements regarding the wave function ψ of an unbound
state in the asymptotic region. We let A exp(ikx)+B exp(−ikx) be the expression
of the wave function ψ in one of the asymptotic regions, for −∞ case.
The total flux of particles when passing a given point is the difference be-
tween the flux (~k/m)|A|2 of particles traveling in the positive sense, and the flux
(~k/m)|B|2 of particles traveling in the negative sense. This flux is equal, to within
a constant, to the Wronskian W (ψ, ψ ∗ ) [20]:

~k  2 i ~k
|A| − |B|2 = W (ψ, ψ ∗ ) (2.39)
m 2 m
The equality of the Wronskian W (ψ, ψ ∗ ) at both ends of the interval (−∞, +∞),
implies that the number of particles entering the interaction region per unit time
is equal to the number which leave it. In accordance with this interpretation, one
or the other of equation (2.35) and (2.36) can be written as:

incident flux − reflected flux = transmitted flux. (2.40)

Considering the same interpretation, we now can define the transmission co-
efficient (transmission probability) T as follows:

transmitted flux
T = . (2.41)
incident flux
One dimensional scattering problems. . . 89

We have in particular
k−∞ k+∞
T+ = |t+∞ |2 , T− = |t−∞ |2 . (2.42)
k+∞ k−∞
This result again shows that the absolute values of the two sides of equation (2.37)
are equal, and one again obtains the equality

T− = T+ . (2.43)

Thus the transmission coefficient of a wave at a given energy is independent of the

direction of travel. This is the reciprocity property of the transmission coefficient.
It is just as hard to traverse a potential barrier in one direction as in the other.
The equality of the absolute values of the two ways of representing the Wron-
skian appearing in equation (2.38), coupled with the conservation relations (2.35)
and (2.36), again yields the reciprocity relation (2.41), and we also obtain relations
between the phases of the reflection and transmission amplitudes:

phase(t+ ) = phase(t− );
   
r+ r−
phase = π − phase .
t+ t−
The most interesting point for these relations is the fact that the phases are related
to “retardation” effects in the propagation of the wave packets, with equivalent
relations for α and
p β. As previously, we can re-scale r and t by absorbing appro-
priate factors of k±∞ , and so simplify the discussion as in the previous section.
(We will not repeat the details of the analysis, as it is straightforward.)

3 Bogoliubov transformation
To see why the Bogoliubov transformation is important, and how it relates to
the transmission and reflection amplitudes, let us consider the canonical commu-
tation relation for bosonic creation and annihilation operators

[â, ↠] = 1. (3.1)

Define a new pair of operators

b̂ = u â + v ↠; (3.2)
† ∗ † ∗
b̂ = u â + v â; (3.3)

where the equation (3.3) is the hermitian conjugate of the equation (3.2). This
transformation is a canonical transformation of these operators. It is easy to find
the implied constraints on the constants u and v. For instance, if the transforma-
tion remains canonical, then by expanding the commutator we see

[b̂, b̂† ] = [u â + v ↠, u∗ ↠+ v ∗ â] = |u|2 − |v|2 [â, ↠].


(3.4)
90 P. Boonserm and M. Visser

Therefore, it can be seen that

|u|2 − |v|2 = 1 (3.5)

is the condition for which the transformation is canonical. Note that since the
form of this condition is reminiscent of the hyperbolic identity

cosh2 r − sinh2 r = 1, (3.6)

between cosh and sinh, the constants u and v are usually parameterized as

u = exp(iθ) cosh r; (3.7)

v = exp(iθ) sinh r. (3.8)

4 Transfer matrix representation

We can also investigate quantum mechanical tunneling by the so-called “trans-
fer matrix method” or “transfer matrix representation”. Ultimately, of course, this
is still equivalent to extracting the transmission coefficient from the solution to the
one-dimensional, time-independent Schrödinger equation. As before, the transmis-
sion coefficient is the ratio of the flux of particles that penetrate a potential barrier
to the flux of particles incident on the barrier. It is related to the probability that
tunneling will occur [33]. We again consider a one-dimensional problem which is
characterized by an incident beam of particles that is either transmitted or re-
flected as a result of scattering from an object. For current purposes it is easiest
to work with potentials of compact support, where V (x) = 0 except in some finite
region [a, b].
As long as the potential V (x) is of compact support, it splits the space in
three parts (x < a, x ∈ [a, b], and x > b). In both (−∞, a] and [b, ∞) the
potential energy is zero. Moreover, in each of these two regions the solution of the
Schrödinger equation can be presented as a superposition of exponentials by

ψL (x) = Ar exp(ikx) + Al exp(−ikx) , x < a, and (4.1)

ψR (x) = Br exp(ikx) + Bl exp(−ikx) , x > b, (4.2)
√
where Al/r and Bl/r are at this stage unspecified, and k = 2mE/~. But because
ψL and ψR are solutions to the Schrödinger equation that can be extended to the
entire real line, and because the Schrödinger equation is a second-order differential
equation so that its solution space is two-dimensional, there must be some linear
relation between the coefficients appearing in ψL and ψR — specifically, there must
be a 2 × 2 matrix M such that
   
Bl Al
=M . (4.3)
Br Ar

The 2 × 2 matrix M depends, in a complicated way, on the potential V (x) in

the region [a, b]. In the transfer matrix approach we shall seek to extract as
One dimensional scattering problems. . . 91

much information as possible without explicitly calculating M . To now derive

amplitudes for reflection and transmission for incidence from the left, we put Ar =
1 (incoming particles), Al = r (reflection), Bl = 0 (no incoming particle from the
right) and Br = t (transmission) in equations (4.1) and (4.2). Then
ψL (x) = exp(ikx) + rL exp(−ikx) , (4.4)
where rL is the left-moving reflection amplitude and on the right of the potential
ψR (x) = tL exp(ikx). (4.5)
where tL is the left-moving transmission amplitude. This tells us that
   
tL 1
=M . (4.6)
0 rL
But since the Schrödinger equation (2.1) is real, the complex conjugate of any
solution is also a solution. Therefore the solution which on the left has the form
∗
ψL = exp(−ikx) + rL exp(+ikx) , (4.7)
must on the right have the form
ψR (x) = t∗L exp(−ikx) , (4.8)
and so we also have    ∗ 
0 rL
= M . (4.9)
t∗L 1
These two matrix equations now imply
∗
 
1 tL −tL rL
M= . (4.10)
1 − rL∗r
L −t∗L rL t∗L
But by conservation of flux we must have
|tL |2 + |rL |2 = 1. (4.11)
We just have seen an important connection between reflection and transmission
amplitudes. In addition, it is also interesting to show how to derive the above
equation by the following argument. From the equation (4.4), we can see that this
corresponds to a flux in the positive x direction. For x < a this is of magnitude
∂ψ ∂ψ ∗
 
~
J = ψ∗ − ψ ,
2mi ∂x ∂x

~ ∗



= exp(−ikx) + rL exp(+ikx) × ik exp(ikx) − rL ik exp(−ikx)
2mi

−(complex conjugate) ,
 
~
= 2ik − 2ik|rL |2 ,
2mi
 
~k
= 1 − |rL |2 . (4.12)
m
92 P. Boonserm and M. Visser

In contrast, for x > b we similarly derive from equation (4.5) the fact that we can
write the flux as

~
t∗L exp(−ikx) × ik(tL exp(ikx))


J =
2mi

∗


− tL exp(ikx) × −ik(tL exp(−ikx)) ,
 
~ 2 2
= ik|tL | + ik|tL | ,
2mi
 
~k 2
= |tL | . (4.13)
m

The probability current J of the wave function ψ(x) is defined as

∂ψ ∗
 
~ ∗ ∂ψ
J= ψ − ψ , (4.14)
2mi ∂x ∂x

in the position basis and satisfies the quantum mechanical continuity equation

∂ ∂
ρ(x, t) + J(x, t) = 0 , (4.15)
∂t ∂x
where ρ(x, t) is probability density. Since there is no time dependence in the
problem, the conservation law in equation (4.15) implies that J(x) is independent
of x. Hence the flux on the left must be equal to the flux on the right, that is, we
expect that
   
~k 2 ~k 2
1 − |rL | = |tL | .
m m
1 − |rL |2 = |tL |2 .

therefore
|tL |2 + |rL |2 = 1 , (4.16)
so
1 1 1
1 − rL∗ r = 1 − |r |2 = |t |2 . (4.17)
L L L

Finally we see that the transfer matrix can be explicitly represented in the form
∗
1/t∗L ∗ ∗
   
1 tL −tL rL −rL /tL
M= = . (4.18)
|tL |2 −t∗L rL t∗L −rL /tL 1/tL

Similarly, we now consider a wave moving in from the right

exp(−ikx), (4.19)
One dimensional scattering problems. . . 93

which then hits the potential, is partially reflected and partially transmitted. In
this case, on the right of the potential we have

ψR (x) = exp(−ikx) + rR exp(+ikx) , (4.20)

where rR is the right-moving reflection amplitude and on the left of the potential

ψL (x) = tR exp(−ikx) , (4.21)

where tR is the left-moving transmission amplitude. This tells us that

   
rR 0
=M . (4.22)
1 tR

Again, since the Schrödinger equation is real, the complex conjugate of any solution
is also a solution. Therefore a related interesting solution which on the left can be
cast in the form
ψL (x) = t∗R exp(+ikx) , (4.23)
must on the right have the form
∗
ψR (x) = exp(+ikx) + rR exp(−ikx) , (4.24)

whence
t∗R
   
1
∗ =M . (4.25)
rR 0
But now these two matrix equations imply

1/t∗R rR /tR
 
M= ∗ . (4.26)
rR /t∗R 1/tR

Combining the information from left moving and right moving cases we have first
that
tL = tR . (4.27)
So we again derive the equality of the transmission amplitudes.

Similarly we see that

rR r∗
= − ∗L , (4.28)
tR tL
implying
∗ tL
rR = −rL ; |rR | = |rL |. (4.29)
t∗L
Note that we cannot in general deduce rL = rR . Indeed, in general this is false.
So for any potential (regardless of whether or not it possesses parity symmetry)
we have
T = |tL |2 = |tR |2 ; R = |rL |2 = |rR |2 , (4.30)
94 P. Boonserm and M. Visser

implying (in the same manner as the previous argument) that the transmission and
reflection coefficients are independent on whether or not the particle is incident
from the left or the right — and we have very carefully not made any assumption
here about any symmetry for the potential V (x) itself. We conclude
1/t∗ −rL ∗ ∗
1/t∗ rR /t
   
/t
M= = ∗ . (4.31)
−rL /t 1/t rR /t∗ 1/t
Note the key step in this general derivation: In any region where the potential
is zero we simply need to solve
~2 d2
− ψ(x) = E ψ(x), (4.32)
2m dx2
for which the two independent solutions are
√
2mE
exp(±ikx); k= , (4.33)
~
or more explicitly  √ 
2mE
exp ± i x . (4.34)
~
To the left of the potential we have
ψL (x) = a exp(ikx) + b exp(−ikx) , (4.35)
while to the right of the potential we have
ψR (x) = c exp(ikx) + d exp(−ikx). (4.36)
Even without knowing anything more about the potential V (x), the linearity of
the Schrödinger ODE guarantees that there will be some 2 × 2 transfer matrix M
such that    
c a
=M . (4.37)
d b
This transfer matrix relates the situation to the left of the potential with the
wave-function to the right of the potential. We could now use this formalism, for
instance, to think about the propagation of electrons down a wire (approximately
one-dimensional) with V (x) used to describe various barriers placed in the path
of the electron. Similar matrices also occur in optics, where they are referred to
as “Jones matrices”.
The components of the transfer matrix M will be some complicated nonlinear
function of the potential V (x), but by linearity of the Schrödinger ODE these
matrix components must be independent of the parameters a, b, c, and d. In some
particularly simple situations we may be able to calculate the matrix M explicitly,
but in general it can only be approximated or bounded [1, 2, 3, 4, 5, 6, 7]. From
the above discussion we now understand, from several different points of view, the
basic concepts of transmission and reflection amplitudes. The probability that a
given incident particle is reflected is called the “reflection coefficient”, R = |r|2 .
While the probability that it is transmitted is called the “transmission coefficient”,
T = |t|2 .
One dimensional scattering problems. . . 95

5 Discussion
In this article, we have presented basic aspects of scattering theory in one
dimension in (we hope) a pedagogically clear manner. For a one-dimensional
model, only one of the three coordinates of 3-dimensional physical space is explic-
itly involved. Specifically, we considered potentials of compact support, when the
potential V (x) is mathematically zero outside of a finite interval. We have just
seen an important connection between reflection and transmission amplitudes, and
how to derive this relation directly by using scattering theory.
We introduced the probability current to express the reflection and transmis-
sion coefficients. The probability current is based on the axiom that the intensity of
a beam is the product of the speed of its particles and their linear number density.
It is then a mathematical theorem that this probability current is conserved. We
then introduced important ideas of reflection and transmission of waves, and have
seen that in principle they are completely specified by the potential function V (x).
For instance, the linearity of the Schrödinger ODE guarantees that there will be
some 2 × 2 transfer matrix. Moreover, this transfer matrix can be represented
by investigating quantum mechanical tunneling by extracting the transmission co-
efficient from the solution to the one-dimensional, time-independent Schrödinger
equation. This general formalism has served as a backdrop for our further inves-
tigations reported in references [1, 2, 3, 4, 5, 6, 7].

Acknowledgement : This research was supported by the Marsden Fund admin-

istered by the Royal Society of New Zealand. PB was additionally supported by
a scholarship from the Royal Government of Thailand.

References
[1] M. Visser, “Some general bounds for 1-D scattering”, Phys. Rev. A 59 (1999)
427–438 [arXiv: quant-ph/9901030].

[2] P. Boonserm and M. Visser, “Bounding the Bogoliubov coefficients”, Annals

of Physics 323 (2008) 27792798 [arXiv: quant-ph/0801.0610].

[3] P. Boonserm and M. Visser, “Bounding the greybody factors for Schwarz-
schild black holes”, Phys. Rev. D 78, 101502 (2008) [Rapid Communications],
[arXiv:gr-qc/0806.2209].

[4] P. Boonserm and M. Visser, Transmission probabilities and the Miller–Good

transformation, Journal of Physics A: Mathematical and Theoretical 42
(2009) 045301 [arXiv: math-ph/0808.2516].

[5] P. Boonserm and M. Visser, “Analytic bounds on transmission probabilities”,

Ann. Phys. (2010), doi:10.1016/j.aop.2010.02.005, 2009, arXiv:0901.0944 [gr-
qc].
96 P. Boonserm and M. Visser

[6] P. Boonserm and M. Visser, “Reformulating the Schrodinger equation as a

Shabat-Zakharov system,”Journal of Mathematical Physics 51 (2010) 022105,
arXiv:0910.2600 [math-ph].
[7] P. Boonserm, Rigorous bounds on Transmission, Reflection, and Bogoliubov
coefficients, (PhD thesis), arXiv:0907.0045 [math-ph]
[8] P. V. Landshoff and A. Metherell, Simple quantum physics, (University Press,
Cambridge, 1997).
[9] L. D. Landau and E.M. Lifshitz, Quantum Mechanics: Non-relativistic theory,
(Pergamon, New York, 1977).
[10] G. Baym, Lectures on Quantum Mechanics, (Benjamin, New York, 1969).
[11] S. Gasiorowicz, Quantum Physics, (Wiley, New York, 1996).
[12] A. Z. Capri, Non-relativistic Quantum Mechanics, (Benjamin-Cummings,
Menlo Park, California, 1985). See esp. pp. 95-109.
[13] P. Stehle, Quantum Mechanics, (Holden-Day, San Francisco, 1996). See esp.
pp. 57-60.
[14] L. I. Schiff, Quantum Mechanics, (McGraw-Hill, New York, 1955).
[15] C. Cohen-Tannoudji, B. Dui, and F. Laloë, Quantum Mechanics, (Wiley, New
York, 1977).
[16] A. Galindo and P. Pascual, Quantum Mechanics I, (Springer-Verlag, Berlin,
1990).
[17] D. Park, Introduction to the Quantum Theory, (McGraw-Hill, New York,
1974).
[18] A. T. Fromhold, Quantum mechanics for applied physics and engineering,
(Academic, New York, 1981).
[19] M. Scharff, Elementary Quantum Mechanics, (Wiley, London, 1969).
[20] A. Messiah, Quantum Mechanics, (North-Holland, Amsterdam, 1958).
[21] E. Merzbacher, Quantum Mechanics, (Wiley, New York, 1965).
[22] J. Singh, Quantum Mechanics: Fundamentals and applications to technology,
(Wiley, New York, 1997).
[23] P. M. Mathews and K. Venkatesan, A textbook of Quantum Mechanics,
(McGraw-Hill, New York, 1978).
[24] R. G. Newton, Scattering Theory of Waves and Particles, (McGraw–Hill, New
York, 1965).
[25] R. G. Newton, Inverse Schrodinger Scattering in Three Dimensions,
(Springer, New York, 1990).
One dimensional scattering problems. . . 97

[26] K. Chadan and P. C. Sabatier, Inverse problems in quantum scattering theory,

(Springer-Verlag, New York, 1989).
[27] W. Eckhaus and A. Van Harten, The Inverse Scattering Transformation and
the theory of Solitons, (North-Holland, Amsterdam, 1981).
[28] N. Froman and P. O. Froman, JWKB Approximation: Contributions to the
Theory, (North-Holland, Amsterdam, 1965).
[29] N. Froman and P. O. Froman, Phase-integral Method: Allowing Nearlying
Transition Points, (Springer, New York, 1996).
[30] N. Froman and P. O. Froman, Physical Problems Solved by the Phase-Integral
Method (Cambridge University Press, Cambridge, 2005).
[31] R. Peierls, Surprises in Theoretical Physics, (Princeton, Princeton, 1979). See
esp. pp. 21-22.
[32] S. Korasani and A. Adibi, Analytical solution of linear ordinary differential
equations by a differential transfer matrix method, Electronic Journal of Dif-
ferential Equations 79 (2003) 1–18.
[33] Transfer matrix techniques are discussed, at varying levels of detail, in the
textbooks by Merzbacher [21], Singh [22], and Mathews and Venkatesan [23],
and also in the research article by Khorasani and Adibi [32].

Petarpa Boonserm
Department of Mathematics,
Faculty of Science,
Chulalongkorn University,
Phayathai Rd., Pathumwan, Bangkok 10330, THAILAND.
e-mail : [email protected]

Matt Visser
School of Mathematics, Statistics, and Operations Research,
Victoria University of Wellington, New Zealand.
e-mail :[email protected]