Apêndice
Apêndice
Apêndice
This brief section is included only to help you understand Eq. (1.8.5) in the
main text and is by no means comprehensive.
Consider the inversion of a 3 x 3 matrix
(A.l.l)
The elements of Af have been named in this way rather than as lvfu, for in the
following discussion we will treat the rows as components of the vectors A, B, and
C, i.e., in the notation of vector analysis (which we will follow in this section).
AR=BxC
BR=CX A (A.l.2)
CR=A X B 655
656 which are said to be reciprocal to A, B, and C. In general,
APPENDIX
and cyclic permutations (A.l.3)
(bR)l
(bRh
(bRh
then
(A.l.4)
where the last equality follows from the fact that the cross product may be written
as a determinant:
(A.l.6)
(We shall follow the convention of using two vertical lines to denote a determinant.)
Hence the inverse of the matrix M is given by
fv!_
lvr' __ (A.l.7)
det i\1
When does det l.J vanish? If one of the vectors, say C, is a linear combination
of the other two; for if
C=aA+/1:8
then
(A.1.8)
The resulting vector is orthogonal to the ones in the product, changes sign when we
interchange any two of the adjacent ones, and so on, just like its three-dimensional
counterpart. If we have a matrix M, whose n rows may be identified with n vectors,
At, A2 , ••• , An, then the cofactor transpose has as its columns the reciprocal vectors
AtR, ... , AnR, where
(A.l.9)
One tricky point: the cross product is defined to be orthogonal to the vectors
in the product with respect to an inner product
and not
A·B="
L... A*B I I
even when the components of A are complex. There is no contradiction here, for
the vectors At, ... , An are fictitious objects that enter a mnemonic and not the
elements of the space '~.r( C) on which the operator acts.
~]-! =[ ~ =~
1 -1 3 -2
!]
and
[~0 -1
~]-!=_!_[-:
2 12
-~
4 -2
-821]
658 Theorem A.l.l. If fll V) = 10) implies IV)= 10) then g-t exists.
APPENDIX
Proof Let I Vt), ... , IVn) be aLI basis in ·~r. Then another LI basis is gener-
ated by the action of n, i.e., fll Vt ), ... , fll Vn) is also a LI basis. To see this, let us
assume the contrary, that there exists a relation of the form
with not all a;=O. Upon pulling out n, because it is linear, we get
with not all a;= 0, which is not true. So we can conclude that every vector IV') in
~;n may be written as a unique linear combination in the new basis generated by n
as
In terms of IV)= 2: a;l V;), we see that every IV') in ' 1.-,;n may be written as
I V')=fll V)
where IV) is unique. In other words, we can think of every IV') in 'l,;n as arising
from a unique source I V) in ~;n under the action of n. Define an operator A whose
action on any vector IV') in ~;n is to take it back to its unique source IV). (If the
source of IV') were not unique-say, because there are two vectors I Vt) and I V2 )
that are mapped into IV') by n-then we could not define A, for acting on IV'), it
would not know whether to give I Vt) or I V2 ).) The action of A is then
A=n- 1 or An=/
(
lo a)= f e-ax dx,
2
a>O (A.2.1)
-x
l~(a)=f·.-c 1 2
" eap'pdpd¢
o Jo
= rr/a
Therefore
By differentiating with respect to a we can get all the integrals of the form
For example,
dx
(A.2.3)
The integrals l2n+l(a) vanish because these are integrals of odd functions over an
even interval -oo to +oo. Equations (A.2.2) and (A.2.3) are valid even if a is purely
imaginary.
Consider next
J
,.. --:~_ , '-I
lo(a, /3) = 4a e -
a(Y ··{3 · 'a)'
- c.x · (TC)
j = e f3'•4a ---- (A.2.5)
----:r \a/
These results are valid even if a and f3 arc complex, provided Re a> 0. Finally. by
applying to both sides of the equation
Jo r x
e
··m d
r=
I
a
f'(z)= I, r · e'
0
1 dr
A complex variable z can be written in terms of two real variables x andy, and
i=(-1) 12 , as
z=x+(v (A.3.l)
One may invert these two equations to express the real and imaginary parts, x and
y, as
You may verify that z = z' implies that x = x' and y = y' by considering the modulus
of z-z'.
From the power-series expansions
It is clear that
eix has unit modulus (x is real).
z=x+iy=(x2+l)1!2l· , x 2
(x··+y)~-
112·-1--i~~.
(x+y)
{nJ
1"
= p(cos 0 + i sin 0)
=pew
where
(A.3.6)
Clearly
lzl =p (A.3. 7)
Each complex number z = x+ iy may be visualized as a point (x, y) in the x-y plane.
This plane is also called the complex z plane.
l 1
--=?f'-±in8(x) (A.4.1)
x"fic: x
662 where s-->0 is a positive infinitesimally small quantity. Consider an integral of the
APPENDIX form
Viewing this as the integral on the real axis of the complex z = x + iy plane, we see
that the integrand has an explicit pole at z = is in addition to any singularities f
might have. We assume/has no singularities on or infinitesimally close to the real
axis. As long as s is fixed, there is no problem with the integral. For example, iff
has some poles in the upper half-plane and vanishes fast enough to permit our closing
the contour in the upper half-plane, the integral equals 2;ri times the sum of the
residues of the poles off and the pole at z =is. Likewise, if we change the sign of
the s term, we simply drop the contribution from the explicit pole, which is now in
the lower half-plane.
What if S-->0? Now the pole is going to ram (from above) into our contour which
runs along the x-axis. So we prepare for this as follows. Since the only singularity near
the real axis is the explicit pole as z =is, we make the following deformation of the
contour without changing the value of I: the contour runs along the real axis from
-oo to -s', (s' is another positive infinitesimal) goes around counterclockwise, below
the origin in a semicircle of radius s', and resumes along the real axis from x = s' to
oo. The nice thing is that we can now set e = 0, which brings the pole to the origin.
The three parts of the integration contour contribute as follows:
The sum of the two integrals in the limit s' -->0 is defined as the principal value integral
denoted by the symbol f/J. In the last term, which is restricted to the infinitesimal
neighbourhood of the origin, we have set the argument of the smooth function f to
zero and done the integral of dzjz counterclockwise around the semicircle to get i;r.
Eq. (A.4.1) is a compact way to say all this. It is understood that Eq. (A.4.1)
is to be used inside an integral only and that inside an integral the factor lj(x-is)
leads to two terms: the first, f!J(ljx), leads to the principal value integral, and the
second, i;ro(x), leads to i;rf(O).
It is clear that if we reverse the sign of the s term, we change the sign of the delta
function since the semicircle now goes around the pole in the clockwise direction. The
principal part is not sensitive to this change of direction and is unaffected.
It is clear that if we replace x by x- a the pole moves from the origin to x =a
andf(O) gets replaced by f(a) so that we may write
1
f/J--±i;ro(x-a) (A.4.4)
(x-a)'fie (x-a)
It is clear that the limits on x need not be ±oo for the formula to work.
Finally, note that according to Eq. (A.4.4) the difference between the integrals 663
with two signs of s is just 2Jrif(a). This too agrees with the present analysis in APPENDIX
terms of the integral I in Eq. (A.4.2) since in the difference of the two integrals the
contribution along the real axis cancels due to opposite directions of travel except
for the part near the pole where the difference of the two semicircles (one going
above and going below the pole) is a circle around the pole.
Answers to Selected Exercises
Chapter 1
1.8.10. w=0,0,2;}c=2,3,-l.
Chapter 4
4.2.1. (I) I, 0, -1
1/2 ]
(3) !Lx= 1)--> [ 1/2 112 ,
1/2
l/2 ]
ILx=-1)-+ [ -1/2 112
1/2 665
666 (4) P(Lx=l)=J/4, P(Lx= 0) = 1 P(Lx=-1)= 1/4
ANSWERS TO
SELECTED
EXERCISES
(5) Ill')
(l/4+ l/2)
[
1
/~ ] =projection of I !Jf) on the
1/21/2 __
L; = 1 eigen-
Chapter 5
(b) T= ( cosh 2 2Ka + a 2 sinh 2 2Ka) ---I where iK is the complex wave number
for !xi sa and a={V0 ----2E)/[4E(V0 --- E)f 12 .
Chapter 7
(2) (X(t)) = (n/2mw) 112 cos wt, (P(I)) = -(mwft/2) 112 sin wt
(3) (X(t)) = (in)- 1([X, H]) = (P(t)) /m, (P(t)) = -mw 2 (X(t) ). By elimin-
ating (P) we can get an equation for (X(t)) and vice versa and solve it
using the initial values (X(O)) and (P(O) ), e.g., (X(t)) =
(X(O)) cos wt+ [(P(O))/mw] sin wt.
Chapter 10 667
ANSWERS TO
10.3.2. r 112 [1334) + 1343) + 1433)] SELECTED
EXERCISES
Chapter 12
Chapter 13
13.3.2. Roughly 1 A.
Chapter 14
14.3.5. (a+ 8)
M= - 2 - (!3 +2 -r) CTx+l·(/3-
I+ - -s- r) CTy+ (a-
- 8)
2 - CT=
14.5.4. ( 1 +c2os 0) 2
Chapter 15
l- (2L·S )/Fi 2
15.2.6. [p- = -------
21+ 1
Chapter 16
16.1.3. -ma5Vb/n1f
Mc = 938.28 MeV
2 (21-f is the proton mass)
a 0 =fz 2 /me 2 =0.511 A
efz/2mc = 0.58 X w--s eV /G (Bohr magneton)
k=8.62x 10-5 eV/K
kT~ 1/40 eV at T= 300 K (room temperature)
1 eV=l.6x I0- 12 erg
vjc=.=f3=a
669
Index
Hamiltonian formulation, 86
Hamilton's equations, 88 Ket, 3
Harmonic oscillator Klein~ Gordon equation, 564