5.
73 Lecture #2
2-1
free particle V(x)=V0
Last Time:
general solution
= Aeikx + Beikx
A,B are complex constants, determined by boundary conditions
k=
p
h
(from e ikx , eigenfunction of p, and the real number, p, is the eigenvalue)
1/2
2m
k = (E V0 ) 2
probability
for E V0
P( x ) = * = |1
A4
|2 2
+4
| B4
|2 + 21Re
( A * B) cos 2 kx2
+ 4444444
2 Im( A * B)sin 23
kx
4
3
4444444
const.
distribution
wiggly
only get wiggly stuff when 2 or
more different values of k are
superimposed. In this special
case we had +k and k.
TODAY
and
1. infinite box
2. (x) well
3. (x) barrier
and
5.73 Lecture #2
2-2
What do we know about (x) for physically realistic V(x)?
( ) = ?
* ( x) ( x) for all x?
* ( x) ( x)dx ?
Continuity of and d /dx ?
Computationally convenient potentials have steps and flat regions.
infinite step
finite step
infinitely high but infinitely thin step, -function
continuous
d d 2
,
not continuous for infinite step, and not for -function
dx dx 2
d
is continuous for finite step
dx
More warm up exercises
1.
Infinite box
V(x)
0
( x) = Ae ikx + Be ikx = C cos kx + D sin kx
[C=A+B, D=iA iB]
(0) = 0 C = 0
(L) = 0 kL = n
n = 1, 2 ,
( why not n
= 0?)
5.73 Lecture #2
recall
2-3
2m n 2 2
= 2
V0 = 0
L
h2
Insert kL = n boundary condition.
k 2 = ( E V0 )
En = n2
h2 2
2 h
2
2 = n
mL
2 mL
8
here.
En is integer multiple
of common factor, E1.
Important for
# of bound levels wavepackets!
n = 0 would be
empty box
E1
normalization (P=1 for 1 particle in well)
L
|D|= (2 / L)1/2
1 =| D |2 dx sin 2 ( nx)
0
n ( x) = (2 / L )1/ 2 sin( nx)
because
L
2
0 sin ( nx)dx = L / 2
i
D = (2 / L)1/2 e{
arbitrary
phase
factor
cartoons of n(x): what happens to {n} and {En} if
we move well:
left or right in x?
up or down in E?
Infinite well was easy: 2 boundary conditions plus normalization requirement.
Generalize to stepwise constant potentials: in each V(x)=constant region,
need to know 2 complex coefficients and, if the particle is confined within a
finite range of x, there is quantization of energy.
* boundary and joining conditions
* normalization
* overall phase arbitrariness
So next step is to deal with case where boundary conditions are not so
obvious. (x) well and barrier.
5.73 Lecture #2
2-4
0
V(x)
V(x) = a (x)
a has units Energy x Length
(because, as we will see, (x) has
units of reciprocal length)
a>0
= 0 everywhere except V(0) = a
strength of the -function well
Schrdinger
Equation
Integrate:
2m
d2
(3
E4
+2
a4
x) 2
2 = (1
dx
E V( x ) h
2 mE
2 ma
dx
dx
x
x
x
lim
=
lim
(
)
+
(
)
(
)
h2
0
0
dx2
h2
d
d
=
LHS =
size of discontinuity
dx x =+ dx x =
+
d2
in
d
at x = 0
dx
RHS = 0
because
2mE
(0)
h2
is finite and integral
over region of length
2 0.
2ma
h2
(0)
because, by the definition of a fn
(x)(x)dx = (0)
or, more generally
(x a)(x)dx = (a)
5.73 Lecture #2
2-5
Since the potential has even symmetry wrt
x x, (x) must be even or odd (not a
mixture) with respect to x x, thus (x) = (x). If (x) is even, there must be a
cusp in (x) at x = 0
(x) is
(x)
continuous
0
OR
BUT NOT
(x) is not
continuous
at x = 0
So what happens
when (x)
is an odd function?
d(+) d()
2ma
= 2 (0)
dx
dx
h
The new
boundary condition
since there is + reflection symmetry for an even (x)
d(+)
d()
=
dx
dx
ma
d()
= m 2 (0)
dx
h
Now find the eigenfunctions and eigenvalues. Standard procedure: divide
space into regions and match and d/dx across boundaries.
5.73 Lecture #2
2-6
Region I
Region II
0
L = I = A L e + x + BL e x
(8 unknowns, because A and B can
be complex numbers)
V(x)
|x|>0
E= |E|
Let E < 0
R = II = A R e + x + BR e x
1/ 2
| E | 2m
=
h 2
(THIS IS WHAT WE DO WHEN k
WOULD BE IMAGINARY)
(+
)=0
AR = 0
unknowns
determined
(2)
(
)=0
BL = 0
(2)
L()=
R(+)0
AL = BR A
(2)
arbitrary phase
(1)
normalization
(1)
(8)
d R ( + )
ma
= Ae 0 = 2 A(0)
h
dx
=
ma
h2
d L ()
+ ma
= + Ae +0 = 2 A
(0)
h
dx
again
ma
= 2
h
L = Ae x
R = Ae x
Done!
required discontinuity in d/dx at
x = 0.
5.73 Lecture #2
2-7
Only one acceptable value of one value of E < 0
2 2
2
ma
ma
h
= 2 |E|=
=
= E
2
h
2m
2h
E=
ma
2h 2
Actually, the above solution was specifically for an even (x). What
about odd (x)? No calculation is needed. Why?
Normalization of
1 = | |2 dx
R = Ae max/ h
1 = 2 | A |2 e
ma
A = 2
h
2 ma h 2 x
h2
dx = 2 | A |2
2 ma
see Gaussian
Handout
1/ 2
ma
= 2
h
1/2
e ma|x|/h
only one bound
level, regardless
of magnitude of a
large a, narrower and taller
There is a continuum of s possible for E > 0. Since the particle
is free for E > 0, specific form of must reflect specific problem:
e.g., particle probability incident from x < 0 region. It is even
more interesting to turn this into the simplest of all barrier
scattering problems. See Non-Lecture pp. 2-8, 9, 10.
5.73 Lecture #2
2-8
Nonlecture
Consider instead scattering off V(x) = + (x)
a>0
V(x) = +(x)
x
0
L = A L e ikx + BL e ikx
2mE
k= 2
h
R = A R e ikx + BR e ikx
1/2
In this problem we have flux entering exclusively from left.
The entering probability flux is |AL|2.
Two things can happen:
1.
transmit through barrier
|AR|2
2.
reflect at barrier
|BL|2
There is no way that BR can become different from 0. Why?
2
Our goal is to determine A R and BL vs. E
L(0) = R(0)
continuity of
A L + BL = AR + B R
but BR = 0
A L + BL = AR
2ma
d R (+0) d L (0)
= + 2 (0)
dx
dx
h
2ma
R(0)
ikA R (ikA L ikBL ) = 2 A R
h
AR = AL + BL
2ma
ik( A L + BL ) ik(A L BL ) = 2 ( A L + BL )
h
L(0)
5.73 Lecture #2
2-9
2ma
( A L + BL )
h2
2ma 2ma
BL 2ik 2 = 2 A L
h
h
2ikBL =
h2
2ma ikh 2
AL
=
1
2ik 2 =
BL 2ma
ma
h
+1 =
ikh 2
ma
B
A R = A L + BL = A L L + BL = BL + BL = BL ( + 1)
BL
= AL/BL
ikh 2
A R = BL
ma
Transmission is
T=
AR
AL
Reflection is
R=
BL
AL
2
2
2
2
What is T(E), R(E)?
AR
= BL
k 2h 4
2 2
m a
= BL
2mE h 4
h
2 2
m a
= BL
2h 2 E
ma 2
ikh 2 ikh 2
AL AL
1
1
=
BL BL
ma
ma
k 2h4
2h 2 E + ma 2
2 = 2 2 +1 =
m a
ma 2
BL
AL
2h 2 E
ma 2
R(E) = 2
=
+ 1
2h E + ma 2 ma 2
decreasing to zero as E increases
ma 2
2h 2 E
T(E) = 2
=
+
1
.
2h E + ma 2 2h 2 E
R(E) + T(E) = 1
increasing to one as E increases
5.73 Lecture #2
Note that:
2 - 10
R(E) starts at 1 at E = 0 and goes to 0 at E
T(E) starts at 0 and increases monotonically to 1 as E increases.
Note also that, at E =
ma 2
2h 2
R as E approaches ma2/2h2 from above and
then changes sign as E passes through ma2/2h2 !
This is the energy of the bound state in the (x)-function well
problem.
See CTDL Chapter 1 Problem #3b (page 87) for a
related problem
