OneDimensionalSchidingerEquation

In the trans
chapter we
edit EE E aiqtaei
of motion and we discussed several generalproperties of the
solutions including
normalizibility linearlyindependent set of
possible solution's wilt discrete energies Now we will exactly
solve for few famous examples in ID

AIFreefarlich Freeparticle means there is no potential

So the time independent schrodingerequation
is
Tty Imy Indian it Yen t Exact I

Comparing it with the relativistic wave equation for the electro

magnetic wave photos we see Heat non relativistic particles

have ist order time dependence the general solution remains

the same as
Xp A EX ED Be
in ÉI
k
which of two linearly independentsolutions in
constist

position due to it and order derivative term the wave rector

is easily obtained from end as

E KYI p
this is difference from the relativistic ware er when E ctr
 One also write instead in terms of sin kn cusk'd
may egg
but they are the same solutions

The eigenvalues only depend on k anddoes not

energy

depend on sign of k whereas the wavefunctions for t k k

the

k are
linearly independent This is something new that we
have not talked about before Such solutions are called
degenerate solutions

A B are arbitrary
constants tobe determined Henceforth we will
only focus on the time independent part where Ya D Had éietl

Since E is energy only kinetic energy so k is real and

the solutions e ik and e ikr are always oscillatory in
space ie
always propagating and never decays even as
we go to K I So this is to
going cause problem
a

in the normalizibility condition that all warefunction must

follow Withotloosing generality we only consider one
i ka
particular solution e and make the k variable
say
extend to t k to k then we have a simplerform to
ikr 4
deal with Ya try c e

where c is the constant to be determined I k correspondsto

ware travelling right and left directions
It is interesting to see that Yale is also a simultaneous
eigenfunction wavefunction
of the momentum operator
with momentum eigenvalue fo kk as

PY it fit fo y 5

Dtm commutes
This is obvious also because the Hamiltonian I
with the momentum operator Therefore in these eigenstates
momentum is
fully determined and it has no uncertainty
Ap o So position is completely undetermined Anta
which i expected became the ware is propagating
for ever in space and is not arable so we cannot

define its position Therefore plane waves are alsocalled

momentum eigenstates or momentum status
simply

Ikaw Itidn if dn a
Cs

Recall that this was the reason we had to invent warefacket

which wasto confine the plane waves in a finite size
Practically this means we have to introduce a confinement
at the two boundaries of a length 2 This introduces a

maximum uncertainty in position to be ant L and finite

Ap also The plane wares reflect back atthe boundary and
create standing waves which are He warepackets
 worry
about
ist the confiningpotential
tt
at the
boundary of a box but

only allowed within a lengthscale of l

In reality this shouldbe done by adding a confining potential called

particle in a box which creates warepacket but
for the timebeing
we will continue to use plane waves and only restrict theintegralfrom
o to L C this is clearly not a physical B C but rather a
mathematical treatment

Kyan et
find n t p

So a Fe we will not worry about 1h

phone part since any phone
so
Rft it c can be absorbed in Y due
to its constant
flare freedom
This is called Boxnormalization where o IL We will
see that in the example of particle in a box problem later by

setting v a in which all possible warefunction Yen hereto

vanish at the wall which will quantize the possible valuesof
R and energy H W Estimate R E forthe longestwavelength
care wing uncertainty principle
Periodic.Boundarycordition
a o

We imagine there is peridicity in the system due to

some

atoms molecules sitting at lattice sites in solid statesystems

ett so that the wave function must follow the peridic
boundary condition
The 4 Xen
i kn
e ei kl end c e
i ki I
e

MIEN when i
he 0 Il I 2

and the normalization condition becomes

1 Y dn I ca lie

Now we see that the the feridic boundary condition makes k

discrete in units of 2011 and here the energy is also
2
discrete En no
15
HW Is the momentum still conserved or the wavefunction is

still a simultaneous wave function ofthe momentum operator

Ans Its consered modulo IMI L J
Now we see that k or n are the discrete quantum numbers of
the wavefunction
I et MI
n
Yn n egg

So different values of k or n corresponding to different linearly

independent eigenfunctions except In
status which are degenerate

and the orthogonal condition becomes

EnCn D da
Yn Cn MmCal d n
11 e

Sm n from the definition

o of Kroneckerdelta
th
 int
to the box normalization discussion

with completely rigid wall vs a

gtintd I

µ ma
112 2221 2

Therefore the probability of finding

the particle outside the well is
É
completely zero therefore all possible I
n
wavefunctions must vanish
at a I 42 Became He 1
s
potential inside is zero so the solution

ofthe schrodinger equation is still

plane wanes with the conditionthat I

only those plane waves solutions are allowed which home

nodes at 2 1112 This is to say all the wavelengths
X are integer multiple of L as we may have seen in
the discussions of ware theory in other courses Therefore
the condition
c e
th o
Macy
C o n a trivial solution But there are more non trivial
solutions To find that ont we exploit the freedom of choosing
the wavefunction
linearly dependent and gauge flare
freedom to satisfythe boundary condition In fact as we will
see more and more later even for a given potential ei the
Schrodinger equation of motion the form ofthe solution and
hence quantized energy eigenvalues
mong change if we change
the boundary condition To rewrite the general solution in
a form suitable for our
boundary
conditions we proceedas

i ka t B i ka
Yong A e e

At B as Kx i CA B Sin Rx I

Then 41144 0 CATB Cos KY I i CA B Sin14 0

C Now we have two choices A B sink 0

R It n 2,46
or A B Cos Kt 0

k ME n
113,5
we do not necessarily have to consider the negative n values
since those negative n values are obtained from positive n values

by changingthe sign of A B So they are not linearly

solutions
independent
Therefore we have two sets of linearly independent
solutions for even odd integers of n therefore the
momenta and hence wavelengths are quantized for a
particle in a box and the corresponding quantized

energies are

If
EE.ie 43

Its easy toprove that the eigenstates are orthogonal and

they can be normalized as

I dn Un Um L so we deride the

eigenstates
by la which gives orthonormalized eigenstates

We notice that the alternative eigenstates

Parity 8 being
coskx a sink x are even and odd functions of
actually not a
no This is coincidence It actually
comes from the symmetry
of the Hamiltonian itself
You take all points nd o in the Hamiltonian and invert
a H then yousee that the potential profile remains
unchanged If the Hamiltonian has a symmetry then the

also have the eigesymmetry

same states
In otherwords the eigenstates

ofthe Hamiltonian is also a simultaneous eigenstate oftte

operator representing that symmetry This symmetry is called

parity Lets whosejob

say p is the symmetry operator
is to invert the position variable a to se in the eigenstate
So P is defined as PYn n Yne a XD
I Don't look for any mathematical form of the P operator Hs
an abstract operator whosejob is to invert all n to se then

apply P again on Egly to obtain

P2Yn n P Un e n Yn ng
So Yn n is an eisenstate of p with eigenvalue l since
P is an Hermitian operator whoseeigenvalues are real
so the eigenvalues of P can be Il So we get
P Ya la In Ga I told
where 1 1 eigenvalues correspond to the eren odd
valves of n Therefore owingto the parity symmetry ofthe
Hamiltonian the eigenfunctions are also eigenfunctions of
the parity operator and hence each eigenstates hone well
defined parity either even or odd in spatial inversion
If a Hamiltonian does not hone
parity a spatial the
inversion symmetry then the eigenfunctions also don't
have this symmetry as the eigenstates are not purely odd
or even in spatial inversion but a linear combination

of Item

AI l Assume the potential i now shifted to k e at

Nz 0 I and VCH o for O LKL Does His system

hone parity
2 Estimate the ground state ofthe Hamiltonian
energy
ei E Ift from the uncertainlyprinciple

3
Extend the calculation to 3D find the degeneracy
in each energy levels
c
FinitePotentialwet
Next we will consider a
finite potential well means

negative potential Although

this is more of an idealized

potentialprofile but the conclusions
drown here are imitatively similar to the Yukawa potential
profile attractive potential We are interested in two energy
solutions Edo and Vo E 10 The solution the first care is
of
easy to guess that because of the positive kinetic energy the
solution's

will be plane wore like wilt slight modification dueto the

potential Such solutions are called scattering solution became as
if
a
free particle is being scattered by the potential well although one
might say here that the particle is being attracted towards the
potential well Then the question would be can the particle aviod
the attractive potential and jump to the otherside or become
trapped at the well Charly it depends on the potential height Vo
and width L
The Eco situation is interesting outside the well we

hone negative K E and inside the well positive K E So the solutions

will be decaying outsidethe well classically forbidden region and
oscillatory inside the well Each oscillation with different wavelength
which corresponds to linearly independent solutions will have different
quanta of energy which are the energy eigenvalues
Do larger war length or spread oftte ware packet correspond to
lower energy the other way Its mostly the wave packetwith
or

larger spread uncertainty in position and hence lies spreadin

momenta correspond to the lower states the groundstate
energy
often reaches the optimum uncertainty a pan t One should

not take this condition blindly in the

future courses since for
complicated potential profile and also when potential depends
on momentum like a charge particle in a magnetic field
things can start looking differently although ground state
having optimum uncertainty At Ax it is very muchthe care

in most cases

Because the potential and hence the Hamiltonian has

inversion x e symmetry so all eigenstates will hone

definite parity The ground state war function will hone

one maximum at 2 0 where the probabilitydensity shouldbe
maximum So we can sketch the wave functions in bolt
cones as
follows
 É
E

Ya I th

inftion i

him
Inside
Wim
warepacket
Inside
Aiff
wave packet outside

outside plane ware like decaying solution

can we do the same we can the uncertainty

use

him principle to estimate the

grind state energy
Nino it turn back to
solving the Schrodinger equation
Vo LE 40 case
exactly We first consider the

It TI Vo N n E Nen

If 272 Vo t E ten Lets put

R E
rolo
El
We hone three regions I I II where

potential pens on me mind I

It
and we hone exact solutions At
the boundaries the potential rises

sharply So insteadof salving at the

boundary we will use the boundary
condition of continuity
of wave fraction and its firstderivative
to obtain its value at the boundary

Region I N 42
Vent 0 E El
Region III a 4

If MEI Xin

W Y r 2

Where In
IZMIT
of inverse length so it
K has the dimension associatedwith
decaying
length scale of the ware packet and uncertainty
 So the general solutions are

t kN No i'in
YI CD A e B e
k exponent
Y G c e t D e kn
k
As a a in region I e blows
up So to have normalization
condition satisfied A 0
Similarly D 0 So we have

i
mail.se 7
Qb

Region It 27g Nol ED Yen 0

R Ner
General solution is
when K
Mtf
ik
Ma G E e t F'e ikr
This solution does not have a well defined parity Since
the potential profile is symmetric under inversion so we

anticipate that all solutions will be either even or odd under

parity in fact alternative solutions will be even s odd So
its convenient to write inthe sin ka curlbad form
Teenie 40
For alternating eigenstates E F are expected to be zero
We will evaluate the coefficients B C E F
by the boundary
conditions

C Y C 42 Y C US
B e 42 E costkt F sin
Rt Raj
YI LIL YI 42
Wh
E cos
Ct tf sin E B e Eb

Ii If a 42 Intl news
K B e 1h42 E since t F cos E Ed
e
diet my If mark

E sin E tf cos E KB e 42 sd

For even eigenstates F o Ik f

ktan fay
For odd eigenstate E o
fke kco.tl
1 65

Ears a 65 put constraints on the allowed values of the

the values of k separately forthe
plane ware solutions ie on

even and odd state Typically a plane ware solution has

of k and hence'YÉegy But as a boundary

continuous values

condition is imposed only certain set of discrete k values

 become allowed which for warepacket with uncertainty
in position which is of the order of the width of thepotential
well l In fact the spread of the ware packet is encoded
in the parameter K Notice that the hers the dimension

of Length inverse It roughly gives us a length scale

where the ware packet vanishes outside the well therefore

the t An Tk then gives us the spread of momentum across

its mean wore rector We can see that easilyforthe ground
state

T
The ground state corresponds to large wavelength ei small
waverector R For k to tanCky2 n 1242 Hence from
eq as
k
we get K x 2 Now the K E of the particle is EMILIE
time Now the average momentum of the particle in the
ground state is zero So the
uncertainty in
momentum

Ap Jay LT at h EE Therefore the uncertainty

in position is Aunt EI Since the maximum uncertainty in

the groundstat is of the order of t hence K EE

From here we can also estimate the ground stateenergy
as follows We know that
An
ro e EE EE É Intel FE
therefore 1
7,432
1 4 An T

We will see below that this ground state energy matcheswith

exact calculation they
 instates Our next stop is to solve ears boarded
and find out the values of K and hence
the energy eigenvalues which are the allowedsolutions
Its not
possible to solve ears Ga Qb analytically The idea is at what
valves of energy E the kits I L.lt's match For that we can
plot ktank42 vs E and K vs E and their intersectionpoint
eigenvalues Similarly forthe odd solutions
give the even

From eq a

tan E E KEEFE
FEE
a
WET
whine w
PIE is a dimensionless parameter
lot and
depends on system
particle's mass

We know that tann is a transcendental function defined

k
between My 2 tann 1 21 for n integer and ne 2 at
pot of Rtt S FEI is shown
by blue line Therefore

depending on the valves of W the Rtt S function crosses

only certain positive integer number of fans which gives the
finite number of quantized energy We notice that
ktR.H.
W nk at which puts the
 t

k k

upper limit on the ware rectork We notice that as n so tannen

and MET tann an so e home w a Nit ni both
sides hone the same sign therefore there will always be atleast
one solution no matter how small k This means there will
always eboundstate solution in a potential well suck that
He particle will be confined inside the well and the wavefunction will
die off fast We will see that by going to the narrow potential well

limit below But even for a bound state the wavefunction spreads
outside the potential well and the probability
of finding the
particle outside the well is finite which will not bethe case

for a classical particles

 For odd parity sontions we have the similar result
We see that it is defined
between no to KeDt and
hence it takes the odd
integer rewrite in the previous
if we

form If we make k 42292

the Rtt as misses to cross the
FF
cot442 line and hence no T
Ji
energy
reason
solutions This is the
the odd parity
III
or
got
solution does not give a

ground state solution but it gives the first excited state and
all odd parity excited states

Summery For ELO we have two regions Inside the pot well
K E n positive and we have scattering oscillatory wanewith warerector

12 Outside the well we hare ve k E and hence decaying ware with

decay length É Continuity of wet and and derivative gives a relation

between k k k
saying not all waverectors are possible and that
restriction quantizes energy Small k in larger wavelength always
correspond to lower energy Ground state has highest warelengthand
ist excited state has smaller ware length and so on In all cases

the w of spread out ofthe well implying that the particle has finite
probability of simultaneously being outside the well
 We can take two limits here Ref Marzbecker ch 6

I WidekDeefwed 8 We first consider to a limit

with L
finite This makes Wta on this cane
removing
the R Its FIT On the L H S we have tanx x x as
Cnt t 2
a 2 This quantize k as
RE Get
Therefore we get

Yo Enl Hat Em.ee

M
CtytEEIh
IEnl vo
n 0 2
4
Books write it as as E two n'Kfir where E is
assumed to be ve I have just substituted E I El and

n runs from 1,3 J

Remarkably n o the ground state solution matches well

with the result obtained in eq using the uncertainly
principle
 Ware functions

Infiniteloot
well

Finitepot
well

we see that as we increase Vo the decay length kit decrearis

this means the spread
ofthe wavefunction keeps on decreasing
and the wave function becomes confined inside the potential
well and hence we reproducethe results from our forevions

example
II.Dellafunctionfotential Another limit is to
interesting
take while at the same time L t o such
Vo
waythat
a a

the area under the potential cure I remains finite

to the delta tunction
This actually takes us potential limit as

fi
Vol
Ven Vote g Sta fig
9
watchoutthe
problem with dimension

My p so in the
Ltan Kt ht
jpg
the Ritts

g my
J g son

we hone orebourndstate but

i
no excited status because cot kya line is
not crossed
by the Ert line The bound ground stat
energy can still be estimated from the same uncertaintyprinciple

In the k to limit fromegga K k tan t 1742

m
Now we BE KI MEI ME
So K MY t This gives

If E EE EE
 Thus the attractive ID deltafunction supports onlyonebond
state which is the ground state
energy

W.ec obtanthresfsom full calculations

The schnidinger equation is

44 272 E 8863 414 2

We continueto call region I no region III n 0 whereas

where region II is now a line

In both regions I II we hone V91 O E let
so
ddt 24 y k y B

where K
JEFE corresponding

to the inversere decay length

Solutions in two regions are

k
µ n
He t Be for neo Yaa
k ka
x cry C e
Ye for no Caz
ta
clearly Ae a as k a a and De a as nta

so for the wavefunctions to be normalizable A D haveto vanish

unconfined at no This gives B c

kn
So we art YE n B et fig
But here the difference is that the first derivative of the
wave function in discontinuous a step function Become it
and derivative following the Schrodingerequation has a

singularity divergence at no Sera we know that a

delta function can also be denoted by the derivative
of a step function so we can anticipate that as are remove
a derivative term from the schrodinger equation we will

get the step function so we can write see

Vin gsexy g Ge 1
Then we integrate both sides
ofthe
Schrodinger equation from a to E
E to ta where E is an infitesimal number This gives
d
z ME Et

i T
gsujy EE

From get
eyes we

t
la
1 2B ke

2K Tcg H
therefore from ear we get k 71
Then the ground state energy is

1E 1 39I est which is sameas Ege

 In this

ofthe particle
case the kinetic energy
is always positive
IT
with an increase in kinetic energy in region II Therefore
the problem can be thought of as a scattering problem

particle like behaviorof a plane wave in which the

potential well serves as a scatterer which gives removes

energy ofthe particle

In region IRI we have free particles whose
solutions are
ikr
Malay Aei kn e B e g
C e
K t D e IRA 2
YAI en
Where K
ftp
In region I we have

Egil em MEC Vo E Y CH

K Y G

ik i kn
Y n E e F e G

whine k f2MEÉ
So all three solutions are oscillatory
The interpretations coefficients AB
of the CD EF are as

follows A is the amplitude of 1

an
incoming wane and B is
t
the amplitude the reflected
of
ware from the potential well Ere
I 42 I 42 II
C D follow similarly
We will not be interested in the eigenvalue problem here

Rather the present interest is a scattering

problem We start
with an incoming ware with amplitude A and c is the outgoing

or transmitted from the scattering potential while

ware

B is the amplitude of the reflected wave For such an

initical condition D 0 and the relation between C and A B
will be bridged by E F and the warerectors k k
To figure out the relations
employ the boundary
we

condition ri the continuity of the wavefunction and its derivative

at the two walls

At 2 42
YI C 42 Y C 42
ik 42 kill it 42 Fei 42
A e B et E e yay

DIII y Film in it k
Ae
ik 42
Beira ee Feit4 us
Define Ke Ik I k s OE KIL
 brei Keitt
I th k éitt meth
i se

HW

At n 42
Y 42 Y C42
ik'm ik
E et 42 f e a c e de ikke Gay

It now IFI n 42

E e't f e
ik th
by ceiba perkily
ed
te't meta
f Keitt Keio f
Combining egg and we get

i M21 I I
where
My ME cos All
if sin ki elk fay
Miz ME II single b

where E Mr th y kn th
 As we mentioned we assume the care where a wave is
incident from the left and there is no wonee incident
from right So D 0
Then we get
es

It sina.hn
EI n
Eob

Now recall our definition that A is the incident ware's

amplitude while C B are the transmitted and reflectedwars
Therefore He bansmissiondreflection efficient are

defined as

It R
ft H

FAI show that Tt R l J

T
cosktitsink't

J
it
was

a
E tEsITE
HIIIII as
We notice that the 1 whichis in contradiction to the
expected classical result in which the particle shouldfully

i
I

Clearly transmission starts as El o With increasing energy

the transmission probability oscillates sharply at the beginning with
the maximum value reaching at T l From
eqkaj we see

that maximum transmission occurs when sin k t 0 ie

R l nT N integer
This condition is satisfied when the wave I and the
reflected wave Is are completely out of phase differ by
at phone they have an destructive interference Therefore

the wave can not reflect fromthe wall at 2 02 and the

wave pass throughthe wall This is sometimes called anti

loci in the sense that when particle loosesits

a

momentum via scattering it is said to be localized But here

 due to desortuctive interference no scattering or localization

occurs We can find out the energy levels ofthe well as

E VE Efm gemma ne

n
The minimum transmission occurs when K'la 2

As E ta we reach the classical limit where to

As Vo a the transimission coefficient t o as we can

infer from eq 124 However as the resonancecondition Kent

is reached the potential term exactly vanishes and we hone a

complete transmission
 n
Ven
D Potential Barrier d
i
1
Quantum
Vo
F
I Il I
42 42 n
Next we consider a finite
potential barrier defined
as

Vcu yo yo for 422kHz

by
o otherwise

For El Vo we hone positive kinetic energy at all x and

hence it oscillatory everywhere with less probability Anparticle
of
to be in the finite potential region The solutions are bunch
of plane ware like solutions called scattering solutions We will

not consider this car any further

for 09 E L No we hone three regions I I I as shown
in figure The motivation
of this problem is not to study eisenenergy
normalisibility of eigenstates etc but to demonstrate the quantum
funneling between region III via a potential barrier which
classically forbid tunneling of classicalparticle with re K E

The Schidingerequation in three regions are

dry 2thhE y cry tf read

III a
q
where KK 2 ME
E had
if am
dry
 I
IF EE CE V Yin

K ten when k
2mfÉ yo
since EL Vo
Therefore k is
imaginary
we define
tin when
kf'm g
solutions above two Schrodinger equations
The of the are

easily obtained as
i ka ikr
I Ya n A e B e too k 42
Kat i ka for 42
III Ya G C e D e a
t
I Y Cry E e t F et for z Lil

I
The interpretation of these
solutions are as follows wÉ É

Ée
In region one we have
an incident plane wave

of amplitude Al and a
reflected plane ware from the barrier at se 42 milt

amplitude 1131 The plane wave however becomes

decaying growing solutions inside the barrier with

decay
correspo ding
amplitudes
with
El R IFI respectively Mostly it will
Then the particle ware will
El F1
transmit through the other wall at n 42 and become
again a
plane wave with amplitude CI and some part
of it will be incoming in region II with amplitude Dl

This problem starts looking similar to the care of a

propagating wave is being hit on a transparent glass plane
and we are asking how much light ware pass through
the plate and how much is being reflected absorption

In fact this is indeed the case for the particle's wave nature
and the potential energy no stands forthe transparency
of
the
plate There quantum particle can tunnelthrough a potential
barrier while a classical particle cannotdueto negative K E

Our focus here will be to study an incident particle from

region with amplitude Al how much it transmit to the
I

region III is 41 7 In the classical limit 19 0 But

quantum mechanics allows tunneling emit negative K E Its
clear that the tunneling probability depends on K
which measures the amplitudeof K E and the width ofthe
barrier In fact the parameter K has the dimension of
E
inverse length I and it roughly measures the
decaylength of the wave In other words it roughly
measures howfar the wave can travels before it ceases to
exists ie how far the wave can tunnel for a given
negative K E
 clearly a rough length scale
tunneling to occur
for the
is when the decay length ki a smaller than the width

ofthe barrier die i

µ f y
t

ors
TaffFEj
ht

2M 72
2 y
This suggests even
for a particle ofmassem with E og
it can tunnel a dritance of 2 ofpotential barrier Vo is
Iy
mf H
opacity
This dimensionless quantity 2M
U
14 is called opacity or inversers e
M
transparency

This tunneling is happening because of quantum

nature ofthe particle in due to having an uncertainly
in the rake of its position and momentum Because
of the
uncertainty in position air there a a probability of finding
the particle about An around it mean position this means
the ware function is spread atleast bythis amount of Ise This
uncertainty in position is the cause of the tunneling and
clearly if A n LL then there is a finite probability of
findingthe particle on the otherside of the barrier The
uncertainty Aredefends on the barrier height Vo In fact
the decay length lat is tht measure this uncertainty To of
see that we start with An n
IHo Now the
momentum unertainty at n
Ept tamarin FEET
so we
get
yan.at fTmTuIE5afffromegaD

Associated with the is then

decaylength
there a decay time
ne n AHv n
Izu where u Kk isthe phase velocity group
m
Velocity ofthe particle we out
FEE
en En
HITE where A E LE
 Let us now return to ear and obtain the coefficients A f
by using boundary conditions
The wavefunction and its ist derivative must be continuous
at all positions
At n 42 I KI C 421 VIII 44 G a

dd ED
n yi fefetn.ie
ik i k42 k42 k 42
A e 44 B e E e ee
ik42 i k42 k
ED A e B e
1 Eek 42 f e 421

Define a complex quantity of Lk tile Then we hone

i WL
q e ga e i arL
C I tetra e u
are
c
I ed

similarly at me 42 we match it x and their derivatives

ni WL

Il
or e

Eet t or e
qe
L ai

By combining ears 6 we can get a relation between

A B c D which will tell us how much were function
bass through the barrier and how much a reflected
M
back set
f
2
so we Ag
M 1422
21
i Kl
where Mu Ms cosh k l t Sinh k D e
i
Miz ME Sinh KL

and q LI ka j y In the Ne 4
 FIFI
A B are the incident and
reflected wave's amplitudes
at the n n won um
C
D are the transmitted and
reflected wave's amplitudes at the
Nz 42 wall

We will assume that there is incident wane from the right

gie D 0 Then we get

Iiiigne ca

Then the transmission coefficient is defined as

T 13 lo

it w compute the reflection coefficient at the n 42

wall defined as R IT and show that Rtty
T it
with u

This is the sticking feature of quantum mechanics wave native

that the particle can tunnel or transmit through a potential barrier
Its like if we put two metals where FLIGHT
electrons are free and sandwich betweenthem an insulator
then the election can tunnel between them
If we connect the
two metals with an
ampmeter we will be able to measure
a finite usually very small current The current dimity
that will be transmitted is JE v19 while He current
that will be reflected back is Jr VCA BD where
re TRIM is the phase velocity This is the mechanismused in
the experiment called ScanningTunneling Microscopy where a

metallic tip is scanned slightly above a metallic sample but

He tip is not toned with a sample omg
a potential barrier between the metallictip and
the sample

Typically the average current will bezero and one needs to

provide some bias voltage to give the extra k E er increase
E close to Yo
This is obvious that as E 0 the fransimission to 0 and
T
monotonically increases with E As E t Vo the
transmission cofficient becomes
T
EI C Ej y

This dimensionless quantity

DII
was
defined above
called the opacity or inverse transparency of the barrier
ID In the limit of high bargello and wide large 12
barrier where the opacity is high we have K2 551 and
2 binh KL 226 sh WL a e and we obtain

T I 16 e Kemp
t
161 e very small

ScanmingTinneling
This formula is often used in
Microscopy experiment to fit the data
so the tunneling is
also suppressed exponentially with the lengthscale of Kt as the
wave function dues

2 In another limit of high large Io but narrow smalls barrier

ie higher transparency K 1 41 and we hone

to
IEEE eying cess

sun is small but tint I

tunneling for EL Vo

234Ef
 VN

43 Dellafunction Potentid

We now think of shrinking the

1
width of the above potential barrier L t o
and simultaneously we take Vo t a such that at theproduct
Vol which is the area under the potential barrier remains
finite Lets call this area as g Vol In this limit the
above potential barrier is denoted 8 function
narrow by a

Vol g 8cm
I
Va a

Although a dirac delta function is not a function in the usual

sense its integral represents a valid quantity Such f function
potentials arise as impurity disorder scatterers in solid stale
metals in which the electrons are free but we are askinghow
do they scatter off from point defects or imparity atomsele
Think of a co metallic wine There are defects impurities
in system It electrons were classical particles then they would
have scattered back from impurities and we would have never

obtain any current But thanks to the quantum nature

of
electrons we hone finite funneling of electrons from such

narrow potential barrier delta functions and we obtainfinite

current
 For the delta function potential the condition KL ki is

still obeyed Therefore the transmission coefficient is

given by WH
ftp.t
M I

The current dainty Matt remainsfinite

E Simple Harmonic Oscillator

So have only considered constantpotentials confinedin

far we

a particular
region We will now consider a position dependent
potential Vcrs Unfortunately there are not many potentials that we
can solve exactly and only handful of potential that has exact
solutions For example simple harmonics oscillator singleparticle via
Coulomb potential change particle in a magnetic field We will
solve the first two problems in this course while the last can be

solved using the tricks learned in the first problem and will be
fought in other courses

etng continuous
potential VCD at a minimum can be

approximated by a Harmonic oscillator as follows

Em Elitism
as
VAD
Veno ddtzfn.cnneg
the
No
I 0

dd Inoue nd't Oke nd

att the mini men data o Veno gives an overall shift to the
potentialwhich eventually shifts all the energy values by a
constantvalue this shift does not change the overall result
and does not appear at the wave function This is called the
hero point energy So without looting generality we set
reno o We also shift no O by sint a simple shiftof

reference frame Finally we define

ddTTv
no I k where
k gives the spring constant Then we hone the simple

Harmonic oscillator potential

VCH 12km2 Ey
This in a good approximation to the potential as long as the
energy E ofthesystem is close to Vlad such that the classical
Hunting points food no141 The potential in ear a again
plotted here We will no longer
concern ourselves with the limit on me

In c
nd n
nd
The classical energy is

E LIM kxV Iam et m wit a

when we define a
frequency
w Fm this
gives an
elliptical constant
energy contour on the phone phase but
the take any continuous energy to
particle can e
Shri in the motion of a particle attached to Y
a
string which then oscillates around its
equilibrium positron which we set to be at n o we
have also studied
many particles attached with each
other with springs and then we have seen that there are
normal resonance modes of vibrations in which all particles

vibrate together

Atoms in solid also be modelled

can
by collectionof
small particles of no size attached with each other via a

spring in a periodic manner Their collective vibrations

give similar normal modes w but because the atoms are

small in site and their distances are ai Ite Ao scales their

vibrations exhibit quantum mechanical nature Then these

vibrational waves have particle dual nature which are

called phonon

Here we are only interested in one atomb vibrations To

go from classical to quantum mechanics we need to make

n f as operator which do not comment anymore

Ht fIIm zinnia x

f IIIa Inning y
with Ca pig it
 P
we will
see
contours
iisIia
will now
sina.e.ms
be allowed in
s
hw
only those solutions are normalizable
and physically acceptable then dicrett

energy contours are separated in units of t w as we can

anticefont from our introductory lectures because the phonespace
in decretized dueto uncertainty principle and the smallest area

possible is h So all the energy contours here to be separated

in units of h and w comes on ride bolt dueto dimentional
reason and also to incorporatethe information about Its
potential According to correspondence forinifolis as h so
we should get the classical result which we indeed get

We can first try to estimate the lowest possible energy

which has to be KW occupying Ae lowest possible area of
the plane space from the uncertainlyprinciple In all the previous

examples there was a clear lengthscale in theproblem and

we said the maximum uncertainty in position in thatboyla
scale In the present care the potential is growing
to infininity and there is no obvious lengthscale in the
problem If we fix the total energy E there there in a

lengthscale in He distance between the classical

 turning points where Itt kinetic energy goes to zero Became
the wane function must decay outside the classical
turning point due to negative K E so this length scale can

justify maximum uncertainty in position You will check

a

yoursdoe that the result below a reproducible or not with

this lengthscale Now we will not albume any fixed energy
and obtain the lowestpossible energy
try to
To estimate the uncertainty in resp we need to
hone some idea of how the ground state should looklike
first thing we notice from
egg is that the Hamiltonian is
in variant under n a is its eigenstates hone definite

parity Now the ground stale energy is when the particle spends
most time at the potential minimum Therefore th
frobdility density SCM has a maximum at the potential
minimum re o and has one extremum become the
ground state energy is something which hone the largest
spread of it warefacket So the ground stale ware function
in crew under with this information
parity we can now

estimate Ise Ap A2e TEE and so on

hi

gey
J d n Y CD I 4lb
a
0
because this
this is an odd
1 integral
Ln4 Ian y Cnj Ii yea to
 So An n B Similarly 47 should also be zero
become otherwise the particle will get out
ofAn spring if
it a finite average momentum to be finite Hence at Ekg
Then the expectation value the Mamikonian which gives us
of
the
energy is
day LaFdm t Im HY
InhabitImwvCasey

Interestingly the average energy is determined bythe

spread in position momentum But both and Alo are in Ite
numerator and one could expect the energy is ininimized when
both an o Ab o But that opposite to what we expectfrom
the uncertainly forineifle that if An o then Alo a and vice
versa This would then rather maximize the energy So the

system will make a compromise between Items We have

said that the ground state is obtained when a fan hk
so we substitute A p n tI am in eg

2
E Im gIngs Iz Mwr CID
Then we minimize E with respect to Ar
dE i ti 2
M WAK o
In Im Fagged

cans4 qmh es
 t
so we get Fgroundstale Ink JimI W zm.tw
Ii
It w
This is consistent with the smallest phone space area in both

position momentum as well as in E vs t whine w MTT

in theangular frequency and it is the time ferrod therefore
the ground state is
enemy obtained by
the

time period it takes to complete

rotation's in the smallest Khan space area One
a single
µ h

ask will thenext energy level be obtained

may
by the time it takes to completetwo rotations Shri sometimes

works but not always The being this sort ofreason

quantization via phone space volume in the Bohr Sommerfeld

quantization procedure Wich we categories as old quantum

theory It turns out that not all the integer multiple

solutions
of
n tw h htt a a normalizable and or orthogonal

solutions
of the schnidinger equations Thatwhat we have
seen in other examples where energy quantization's are obtained

in units of n 1st etc Therefore accordingto new

the normalizability and other
quantum Trg boundary
conditions
fats constraint Its possible solutions and that
on

quantises He energy ein raters We will see it again

the simple Harmonic oscillator
for
From ear 5 we now have a lengthscale as the optimum
uncertainly of the ground stole wave function we denote it as

a g defined by a
f w

this same as the distance between Aretwo classical turning

points nd 7 Its defined by the chitances where E Ikki
er where momentum vanishes Kee JIE k
FEW a where we have substituted F LKN L WI JEM
So it indeed the classical turning point that determines
the maximum uncertainty in position
 Interestingly we can solve the schnidinger equations exactly

as a differential equation solution

2h1m DIY I mwr x E re

One
thing we notice is that the potential keeps
growing
forever Therefore whatever the energy is there is a classical

turning point and the wavefunction will decay odride the

classical
turning point The decay length is a E
e Fmtw

Therefore there will always be bound state or localized state

in this system From all the knowledge sofar we eefeftthe
ware function should
Mla in the
decay as e asymptote
limit of se I x We will however see a slightly different
form is a harnessan form e
4th as n Ia this function
also has the variance in
Inside the classical turningpoints
the kinetic energy is position and hence we expect oscillatory
solutions wawpackets wild quantized wavelength

Before we plunged into solving it we will first try to

simplify the look of the above differential equation by

cleaning it up be redefining the fruition variable in
terms
of
a dimensionless variable u

ne ka fmTf n
i
 I This sort of choice of dimensionless variable choice is very
useful in physicsfor book to make the equations look simpler as
well as to be able to put it in a computer But if we want
to solve
eq g it annoying to font the values of f lo
too n
Y
M co 31 etc In fact in most cases these numbers are even
smaller than the smallest members any computer can handle
Therefore choosing a dimensionless variable always gamreenteeto
make it solvable For that purpose we hone one problem ei to
find a lengthscale
which can hide all the
unnecessary
variables Luckily we hone a length scale in our froblem fora
given energy in a
J

Substituting u an in ear we get

HI DWI I moral ritual E Yiu

Q Mar du 2

Meet me substitute a
VIT which gives

dke ut Y
tw
Ture
t I kw E y
z

DIY n y Ew 4 N y
duh

where N w a dimensional number We call it

 N but right now there is no constraint on the possible
values them however does indicate how
Thisform F
of N
the energy eigenvalues are goingto look like The boundary
conditions will put constraints on the allowed values
of n and
hence we will obtain the quantization condition
So the differential ear we hone is

1dI H
Usually we solve differential equations by series method
which gives some finiteforfinite series solution in powers of
U Then as u increases many ofthese series diverges and we

have radius of convergence i e some limits on themaximum

a

value of a ufoto which the series is defined Here we cannot

brt inch sharf cutoff on a scarce Cte potential is monotonically
growing in n We hope there is also exponential partrelated
above the classical
turning point which decays faster than the
power series then the wavefunction will be normazeble
growing
So we first study the asymptotic behavior u I a since
n'in a dimensionless constant it remains constant org at Ia so

the differential emotion can be written as

dry MY dog
fur
 This does not have a powerlaw solution became I As decreases
two powers of while R Ars increases two powers On the otherhand
u

a function of the form N will work became the 2nd order

differential term on the L H S have to generate two powers of n

which exponential of this form can generate So we take an

ansatt as the general solution

u
Tcu flu e
Lib
Wk thats the solution which will
we only consider e became

be normalizable We do expect that tug will be a polynomial

Substituting agus in Leo's we get

daft an data ten D f o

This is atonally as well knowndifferential equator

Its solutions are well known and studied in detailsin
the Math Phys coarse Here we'll focus more on the solutions physical

properties and origin of quantization

First thing we notice isthat the original Schrodinger gratin

is invariant under inversion n s n So the soIntroins must hone
definite parity ii they are erone odd under n i n Now

in e Nk is always eren under u s u Therefore tch must

eq
be even odd also under parity This is also clear from the invariance

of the Hermite equation underparity u u

 We solve it in the standard series solution method by
assuming a

flu E a j UJ B
jaco
where j are positive integers became for negative integer tea
will hone singularity at Ko which we don't want Wesubstitute

eq in CD and collect
for uJ H term as

CtaJCju ajtz Cajei N

aj ud O
fly
T.ge
oL
Non since all hi terms are linearly independent Here
for if
He sum of a series of linearly independent function goestoaero
then every coefficient must vanish this gives us the recursion

relations
aiez
aiC
II.LI
We notice in the recursion relations that became it skips
one coefficient in between we need to set two initial values
ao and a and Hen every other terms are determined This
is not problem because Schrodinger earis 2nd order and
a we

always need to boundary conditions

We also observe that because only even and odd

solutions are allowed
parity Anyone in each ware function
 either all even terms or all odd terms will contribute

but both will not contribute simultaneously In otherwords

when no f o a 0 and we hone even solutions f even

and when do 0
9,70 we hone odd toIntron's fodd
Then essentially have
we
only one free parameter ao or a
which can be determined
by the normalization condition

Bnt there is still a serious foroblem with the general

solutions The infinite series fun does not diverge slower than

the e M term converge We can check the convergence

rate by looking at the ratio between

age and 9J in the

limit j i a which gives

je2
U Ajez I 2J
juggle I 3
U

So the coefficients decrease a II but we hone the W term too

Letsalso check how e WH term converges

k
e
t.EE
idk
Info ht f talk we define 2k I
where j fakes
l IJ hi eroncileser
Eoz
I ai I
j ro 4 same series as
flu
 Then ni Ed
µ
ai ajae IF
clad
12
je uh

Therefore both fay and e Mt converge at the same rate But

growing with increasing
it keeps u therefore the series never

converges as we integrate from a to a in otherwords all

solutions are
not normazable

there is however a hope From ear J that we see

if any of the coefficient a ez becomes 2ero at some nth

term then all subsequent higher coeffients also vanishes Therefore

the server will terminate at a finite n't term From the
recursion relation it clear that it terminates if it takes

interger rakes such that

Dt2 where he o I 2

Recall that the energy is defined as

hw Y
E ihYn n

N I 0 I2

Are normalizibity requirement

Therefore ofthe wavefunction
demands that not all energy eigenvalues are allowed but only

those values which turns out tobe integer multiple

of Kw are
normalizable and hence provides
physical
solutions Shri is how
the quantization of the energy levels arise in modern quantumtheory
the overall of thots is called the Zero point energy and
shift
it makes the ground state energy for nao to be fekete
we can draw those allowed values of the energy onthe
potential profile as Energy

f ray e
izmirror
Eq n 4
I i
E
I n 3

iii I

Xie
F n o Eto
n
Nd n

We see that every energylevels have differentclassical turningpoints

and hence the corresponding wavefunction will hone different spread
or uncertainty ofthe order of nel its
Since the uncertainty is

obvious that the ground state has least uncertainty

Wavefunctions o The corresponding wane function forthe above
eigenvalues
of the following differential equation
are the solution

data an fn o I
941 an

This differential equation called Hermite's differentialequation

a

and its solutions are called Hermite polynomials denoted by

th E An in We can actually evaluate all coefficients an of
fn fromthe recursion relation except the constant ao for the
even valuesof h and as for odd wavefunctions This constant
can be evaluated
by the normalization constant The Hermite
polynomials are actually orthonormalized wilt the weightfactor
e Was defined by
itmay
giant e du 2 n a 8mn

Then the
full wavefunction of the Schrodingereq is

14141454221
1414 where u Ya
and a
Ew
Interestingly the weight factor é turns out to bethe haussian

part ofthe that the wavefunctions are

aut matically
wavefunction so

fat
orthonormalized a
a
casement du Hi 8mn
 where Nn it
So we write the full wavefunction as

t
m
it

H Alternative Hermit
H u 2W polynomials are even

Hz In 4W 2 and odd underfairly

Hz h 842 12 u
as expected

I these Hermit polynomials can alternatively obtained from the

Generating function

I
tan
acu y e p And

We can
plot some of these ware functions
 Alternative wavefunctions are even odd as expected Eren stabs
Lane a maximum at n o and odd States reawiches at no They
are oscillatory inside the corresponding classical turning point
while they decary exponentially outside it To see that we

refolot Are above eigenfunctions on top of the potential profile and

shift each eigenstates vertically info just onlyfor visualization
purpose

i even

I odd

I i even

ood

i even

Xdo XD i

As expected the classicalturning

faint increases with increasing
lends and the corresponding ware furation a oscillatory inside
energy
and decaying astride The number of extremism increases with increasing
 energy lords The spread uncertainty
of the warefunction foreach
energy limb is still determined
by An corresponding ace

HI i show with explicit calculation that the position uncertainty

Ee He variance in position for each warefunction Afsynmatches

with its corresponding classical turningfront nee n

Also compute the
momentum
uncertainty in each state
Do yo think they can be related to the forsition uncertainty

by de Broglie relation Evalue An ad foreach level

1 If we hone a particle with arbitrary energy F which
does not match with
any Moecific eigenenergy of the
Harmonic oscillator but the particle is still attached with
a Mooring think of a vibrating atom which was initially
at some eigenstate then we its temperature inch
increase
that it gains some thermal energy to beliberated fromits

specific energy level How will you exfrees its general

state and how will you evaluate
wave function
of this system
its energy now Explain the physical meaning of
all terms

Iv We change the potential to VCH I kirtan

Then evaluate the eigenvalues and eigenfunctions

Now
fit aninfinite wall at a Ood harmonic potential
UH I knw only for my o and v a for Ko Sketch
the wave function
of this potential profile