Review of Applied Physics (RAP) Volume 1 Issue 1, December 2012

www.seipub.org/rap

TheoryoftheQuantumHallEffect:Cyclotron
Resonance,KohnsTheoremandHubbard
Model
CorrectTheoryofHallEffect
KeshavN.Shrivastava*,
*

SchoolofPhysics,UniversityofHyderabad,Hyderabad500046India

keshav1001@yahoo.com

has unusual quantum Hall effect which has been

explained by means of nonrelativistic theory [14].
Some authors have suggested to use a charge
neutrality point which is located at the center of
negatively charged electron band and the positively
chargedholes.Actually,alloftheparticlesoftheDirac
equation must be negatively charged or positively
charged so that there are no solutions which predict
theelectricallyneutralparticles.Ofcourse,thepointin
betweentwobandscanexistwithouttheexistenceofa
particle. Hence the point in between electron and the
hole bands is called Dirac point. At this point, the
bands apparently cross but a careful calculation [1]
showed that there is a gap of 27 meV in graphene.
Apparently in 1961, Kohn [5] has shown that the
cyclotron resonance is independent of the interaction.
This theorem is of interest to understand the
experimentaldataoncyclotronresonance.Wewishto
understand if this is really true. Some authors [6] are
interested in the dependence of the theorem on the
dispersion relation. Whether the theorem depends on
the k space or due to some reasons such as low
temperatures (<1K), the interaction disappears and
there is no k dependence whatsoever? Usually, the
kinetic energy of the electron has a k2 dependence in
theunperturbedenergy, 2 k 2 / 2m .Withsomestretch
ofimaginationtheenergymaybealinearfunctionof
k.Insuchacasewhetherthetheoremdependsonthe
wave vector dependence of the kinetic energy? We
will see that the correct cyclotron frequency is
independent of the wave vector dependence of the
kinetic energy and hence the Kohn theorem is well
obeyed. We will see that a modification of the
unperturbed Hamiltonian leads to the interpretation
oftheexperimentaldataingraphene.Hence,thedata
does not depend on the wave vector dependence of
thekineticenergyandKohntheoremisobeyedwith
a modification of the unperturbed Hamiltonian. We

Abstract
A monolayer of carbon is called graphene. It exhibits
unusual properties in the Hall effect and in the cyclotron
resonance.Itisfoundthatitexhibitsfractionalchargeinthe
Hall effect. The interactions amongst electrons almost
become constant at low temperatures. Hence, the Kohns
theorem, which shows that the interactions do not play
much roleindetermining the cyclotron resonance, becomes
operativeatlowtemperatures.Theexperimentsongraphene
do not depend on the wave vector dependence of the
frequency. Hence whether the dispersion depends on k2 or
on k does not matter. The Hubbard model has been very
successfulinexplainingthegroundstateofseveralelectron
systems. We consider a triangle with three vortices. Each
vortexcanbeoccupiedbytwoelectrons.Byusingthespinin
aparticularway,wecanobtainnewfeaturesintheHubbard
model. There is a doubling in the PeierlsLuttinger phase
factorandeigenvaluesacquirehighermultiplicitiesthanare
known for the usual treatment of spin. The flux is
distributedontheareaofthetriangle.Thegrapheneconsists
of hexagons of carbon atoms but the Hall effect shows that
there are defects on which electrons form clusters so that
there is spin wave type behaviour. A cluster of electrons
showsspinwavesleadingtospindeviationofseveralper
cent.
This document gives formatting instructions for authors
preparingpapersforpublicationinthisjournal.Allauthors
must follow the instructions given in this document for the
papers to be published. You can use this document as both
aninstructionsetandasatemplateintowhichyoucantype
yourowntext.
Keywords
Graphene;KohnsTheorem;QuantumHalleffect;HubbardModel

Introduction
A monolayer of carbon atoms with hexagonal
structureis called graphene. Thereisno doubt that it

15

www.seipub.org/rap

Review of Applied Physics (RAP) Volume 1 Issue 1, December 2012

willseethatsuchamodificationdependsonspinand
theangularmomentum.

whereNisthetotalnumberofparticles.Thevaluesof
the magnetic field for which levels are filled are
givenby,
1
2e

(12)
B
cA

In this paper, we report the Kohns theorem with a

modification which explains the graphene data and
theactivationenergies.WealsodescribetheHubbard
model for three sites on a triangle. It is found that
PeierlsLuttingerphasefactorsareaffectedbyspin.

where,
A

Kohns Theorem

2 2 N
(13)
L1 L 2

istheinverseareaoftheFermisurface.
Modification. The cyclotron resonance without the
Wefirstdefinethecyclotronfrequencyas,
interactioncanbeobtainedwithfractionalchargesby
eB
c
(1)
a
suitable modification of the unperturbed
mc
Hamiltonian.
In this way the Kohns theorem is
whereeisthechargeoftheelectron,Bisthemagnetic
retained, i.e., we can get the correct cyclotron
field in the z direction, m is the mass of the electron
resonance without the interaction. We replace the
and c is the velocity of light. The momentum of the
equation(1)by,
electronalongydirectionispi,y+(eB/mc)mxi,hencethe
1 eB
componentsare,
c g
(14)

2
mc
Pi [ p i , x , p i , y (eB / mc)mxi , p i , z ] (2)
whereg=(2j+1)/(2l+1)sothat(10)ismodifiedto,
sothattheHamiltonianis,
1
1
En= (n ) c g .(15)
1
N 2
2
2
H=
i 1 Pi U (3)
2m
The equation (15) explains all of the graphene data
For a twodimensional electron gas confined to a
correctly [14] and as we see, there is no interaction
rectangle, L1L2 in the xy plane, the singleparticle
and the Kohns theorem is obeyed in spite of the
Hamiltonianis,
modification.
1
eB
H1=
[ px2 ( p y
mx) 2 ] (4)
Interactions.AsoriginallysuggestedbyKohntheorem,
2m
mc
the interactions play no role in determining the
whoseeigenfunctionsare,
cyclotron frequency. Then what happened to the
n,k exp(iky )u n ( x k / s1 ) (5)
interactions? There are two types of cyclotron
resonances, (i) E E c where E and E are
where
eB
withinoneband.Inthatcase,theinteractionsaregone
(6)
s1
c
inthegvaluewhichdoesnotconstituteaninteraction.
istheinversearea.Eachlevelhasadegeneracy,
Hence, the Kohn theorem is well obeyed. (ii) It is
s1
eB L1 L 2
possible that interactions should be added as an
g 1 L1 L 2
(7)

c
energysothatoneparticleisinthevalencebandand
sothat,
the other is in the conduction band so that the
hc
transition requires an extra energy produced by the
cg 1 eBL1 L2 or g 1 B(2 L1 L2 ) .(8)
interaction. The resonance then occurs at a modified
e
frequency, E E c . As far as the
Hence,g1isanumberequalto1or2.Ingeneral,g1is
aninteger.Thenthharmonicoscillatoreigenfunction
experiments are concerned this will give rise to a
oftheHamiltonian,
modifiedgvalue, E E =(1/2)g .Thegap

1
2
Thenumberoffilledlevelsis .Thehigherlevelsare
empty,

n , k (n ) c .(10)

g1=N(11)

16

is involved in the transition so that the resonance

maybecalledthegapcyclotronresonance.Thistype
of resonances often occur in materials such as GaAs
and they are always accompanied with the cyclotron
resonance. The wave vector dependence of the
dispersion relation does not come into play so that
whetherthefrequencydependsonkoronk2doesnot
matter. The Kohns theorem is thus well obeyed in
graphene[6].

1
H=
( p x 2 2 s1 2 x 2 ) (9)
2m
isunandtheeigenvaluesofn,kare,

Review of Applied Physics (RAP) Volume 1 Issue 1, December 2012

www.seipub.org/rap

Spin Hubbard Model

in the reverse direction. We transform the localized

basis|j>intoaplanewavebasis,

The Hubbard model has been very successful in

explainingthegroundstateofseveralelectronsystems
[7].Alatticeof1212sitescanbesolvedtoobtainthe
ground state energy. It permits the electrons to hop
from one site to another and hence it is exact. When
theCoulombinteractionisadded,theapproximations
are necessary. Some of the models can be solved
exactly when the Coulomb interaction is suitably
decoupled. It is customary to ignore the spin and
calculate the lowest energy levels. Then electrons can
be filled at the rate of one per level and when lowest
levels are complete, two electrons, one with spin up
and the other with spin down can be filled. A lot of
times, this is the only way the spin occurs. We find
that when the spin and the orbital angular momenta
aretakenintoaccount,itispossibletogetdoublingin
theeigenvaluessuchthatfourstatesoccurratherthan
well known two states. There is doubling in the
PeierlsLuttingerphasefactors.Tripledotmolecule.In
this molecule there are three sites andall of the three
sites can be doubly occupied. We consider three dots
locatedonthevorticesofanequilateraltrianglewitha
magnetic field directed perpendicular to the plane of
the triangle. The three vortices are named 1, 2 and 3.
When the electrons jump from site 1 to site 2, they
acquireaphasefactor12.TheareaofthetriangleisA.
Accordingtothefluxquantization,
B. A

| k 3 j 1e ikj | j (21)

where |j>= |1>, |2> or |3> so that the reduced wave

vectorsappearas,
| 1'
| 2'
| 3'

(| 1 | 2 | 3 )

3
1

(| 1 exp(i 2 / 3) | 2 exp(i 4 / 3) | 3 )

3
1

(| 1 exp(2 / 3) | 2 exp(i 4 / 3) | 3 )

(22)
andtheireigenvaluesare,
E1 E 2 | t | cos(2 / 3) g * B H . z S z
E 2 E 2 | t | cos[2 ( 1) / 3] g * B H . z S z
E 3 E 2 | t | cos[2 ( 1) / 3] g * B H . z S z

(23)
where,
1 1 0

(24)
2 0 1

zSz

Usually, this will split E1, E2 and E3 into doublets.

However,inthepresentcase,weconsiderthespinina
more intricate manner so that we get the correct
plateaus in the quantum Hall effect [4,1122]. This
theory explains the Stormers data [2], Pans data [4]
and Tsuis data [21] correctly and all fractions
observed[22,23]intheexperimentsarewellexplained.
Thefluxisquantizedas,

hc (16)
e

Thetotalfluxiswhichisevenlydistributedsothat,
1
3

hc
e (25)

12 (17)

B. A

Similarly,

whichintheHubbardmodelbecomes,
1
3

23 31 . (18)

B. A 312

TheHubbardHamiltonianforthissystemis,

hc
e (26)

Inordertopredicttheplateauscorrectly,wereplaceg
by,

H= E i , ci c i '3 ,i , j 1t ij ( B)c i c j U (19)

1
s
2 j 1
2
g2

2l 1
2l 1 (27)

which is an exact model [1]. Here U represents the

Coulomb interactions. A detailed description of the
HubbardmodelinonedimensionisgivenbyEssleret
al[8].Inthepresentexample,thethreesitesareintwo
dimensions and theHubbard hoppingterm is subject
toaPeierlsLuttingerphasefactor[9,10],

which due to the charge in the Bohr magneton is

equivalent to replacing e by e*=(1/2)ge. The energy
levelsintheabovethensplitas,
1
gB H
2
1
E 2 | t | cos(2 / 3) g B H
2
(28)

E1( ) E 2 | t | cos(2 / 3)

t ij ( B ) t ij exp(2i ij ) .(20)

The current can travel from point 1 to point 2, from

point2topoint3andfrompoint3topoint1andalso

E1( )

17

www.seipub.org/rap

Review of Applied Physics (RAP) Volume 1 Issue 1, December 2012

fortheCoulombinteraction,whetheritappliestothe
data or in what way it enters into the theory? The
repulsive Coulomb interaction between electrons is
givenby,

due to two values of Sz but there are two values of g

duetoj=ls.Here,
1
1
l s
s
2
2
g 2
, g 2
2l 1
2l 1 (29)
l

Hc= ij '

e2
(35)
rij

so that E1(+) has two values and E1() also has two
values. Hence E1 splits into 4 values instead of the
usual2,withsimilarsplittingforE2andE3of(23).

We replace rij by (hc/eB)1/2 so that the Coulomb

potentialbecomes,

Peierls-Luttinger Phase

Thephasefactoris,

12

2l 1 .Thefluxquantizesas,
1
l s
2

(l

1
s )e
2

Forl=0,s=1/2,thereisadivergencesothattheareaof
thequantizedfluxbecomesverylargeforthenegative
sign.For,l=0,s=1/2andpositivesign,
B.A/3=12.hc/e(31)
Forl=1,s=1/2,withpositivesign,
B. A
3 hc
(32)
12
3
2 e

whereasforl=1,s=1/2,andnegativesign,
B. A
hc
12 .3 .(33)
e
3

Similar relations hold for 23 and 31 and many more

canbegeneratedbyvaryingthevalueofl.Weseethat
thePeierlsLuttingerphase[9,10]splitsintotwovalues,
1
s)] (34)
2
forl=0,12=;forl=1,(3/2)12and312etc.
exp(2i ij ) exp[2i ij (2l 1) /(l

Search for an Interaction

Henriksen et al [6] made an effort to search for an
interaction.Itturnedoutthatmostoftheexperiments
carriedoutatlowtemperaturesshowednointeraction
at all. So if an interaction is assumed and compared
with the experiments, it should show that the
interaction is absent. All of the observation can be
explained on the basis of a suitable unperturbed
Hamiltonian. The basic idea of the Kohns theorem is
also that in cyclotron resonance the interaction does
notplayanimportantroleexceptforaconstantsothe
cyclotron resonance can be explained by unperturbed
Hamiltonian.Hencetheproblemisreducedtofinding
the unperturbed Hamiltonian. If we take a bad
formula,itshouldturnouttobeincorrect.Letuslook

18

eB
ij e 2
hc

1/ 2

(36)

whichvariesasB1/2.Usually,thedistancebetweentwo
electrons enters the Coulomb interaction rather than
the distance upto which flux is quantized. Therefore
(36)ismuchsmallerthan(35)buttheadvantageisthat
there is a quantized field instead of the Coulomb
repulsion. If the cyclotron resonance shift is
proportional to the B1/2, then we can conclude that
electronsareconfinedwithintheareainwhichfluxis
quantizedandtherearenoelectronsoutsidethisarea
and there is a reduced Coulomb interaction. If the
experimental data does not give the square root
dependent shift, then it means that there is no
Coulomb interaction as such except that which might
be absorbed in the unperturbed Hamiltonian by a
transformation.Inthecaseofspecialpropertiesofspin
withbothsignsinthetotalangularmomentumj=ls,
there is no interaction and all the interpretation is
based on the unperturbed Hamiltonian [1221].
Similarly, Kohns theorem requires no interaction to
appearin the cyclotron resonance. If the Hamiltonian
is H = gBB.s, there is no interaction and the
unperturbed Hamiltonian is sufficient to explain the
data except that some effect of the interaction is
absorbedinthegvalue.Itwasfound[24]byThomas
in1926,thatthederivativeoftheCoulombinteraction,
Vc, appears in the coupling constant of the spinorbit
interaction,

B.A= 312 hc 312 hc(2l 1) (30)

1
ge
2

R
c

(r )

1 d
Vc (37)
2m c r dr
1

2 2

with
Hso= k (rk ) L k .S k (38)
which is small compared with the Coulomb
interaction, Vc. If the experimental data shows that
shiftisproportionaltoB1/2,thenitmeansthatthereis
interactionsothatKohnstheoremdoesnotapply.On
the other hand, if data did not give B1/2, then there is
nointeractionandKohnstheoremapplies.Theshiftof
the cyclotron resonance measured [6] in a sample of

Review of Applied Physics (RAP) Volume 1 Issue 1, December 2012

www.seipub.org/rap

graphene deposited on Si/SiO2 wafer can fit into a

linear behavior. It is quite clear that data is
insufficient to favor a B1/2 type interaction. The error
bars are very large so that it is difficult to obtain any
definiteconclusionfromthisdata.Therefore,theclaim
to have found the B1/2 dependence of the cyclotron
resonance shift is weak and hence there is no
interaction so we favor Kohns theorem. This result
requires the definition of the interaction. Any part of
theinteractionwhichisdiagonalizableisnotcalledan
interactionwhenitleadstononinteractingparticles.If
a transformation absorbs the interaction into the
unperturbedHamiltonian,thenalsowesaythatthere
isnointeraction.Thedatashowsthattheshiftcanbe
fitted to a linear dependence on magnetic field, B.
Hence, there is no interaction and Kohns theorem is
applicable.Itmaybenotedthat(35)isbetterthan(36).

Henriksen,E.A.,CaddenZimansky,P.,Jiang,Z.,Li,Z.Q.,
Tung,L.C.,Schwartz,M.E.,Takita,M.,Y.J.Wang,Y.
J.,KimP.andStormer,H.L.,Phys.Rev.Lett.104(2010)
067404.
Hubbard,J.,Proc.R.Soc.A240(1957)539;A243(1959)336.
Jiang,Z.,Zhang,Y.,Stormer,H.L.andKimP.Phys.Rev.
Lett.99(2007)106802.
Kohn,W.,Phys.Rev.123(1961)12421244.
Luttinger,J.M.,Phys.Rev.84(1951)814.
Peierls,R.,Z.Phys.80(1933)763.
Radhika Devi V. and Shrivastava, K.N., Spectrochimica
ActaA95,172176(2012).
Shrivastava, K. N., AIP Conf. Proc. Vol. 1017(2008), p. 422
428.
Shrivastava,K.N.,Proc.SPIE(USA)7155(2008)71552F.

Conclusions

Shrivastava,K.N.,AIPConf.Proc.1150(2009)5967.
Shrivastava, K. N., Phys. Lett. A113 (1986) 435; 115 (1986)

We have studied the Kohn theorem and find that the

cyclotron resonance does not depend on the
interaction. It is also found that the wave vector
dependence of the frequency does not affect the
theoremandhenceitiswellobeyedingraphene.We
also find that the PeierlsLuttinger phase splits and
there are multiplicities in the eigen values of the
Hubbardmodelduetotwosignsinthespin.

495(E).
Shrivastava,K.N.,Phys.Lett.A326(2004)469472.
Shrivastava,K.N.,Mod.Phys.Lett.B,13(1999)1087.
Shrivastava,K.N.,Mod.Phys.Lett.B14(2000)1009.
Shrivastava,K.N.,AIPConf.Proc.1169(2009)4854.
Shrivastava,K.N.,AIPConf.Proc.1169(2009)241246.
Shrivastava,K.N.,AIPConf.Proc.1250(2010)2730.

In recent years the picture of thequantum Hall effect

hasbeendescribedcorrectly[2530]anditisclearthat
Laughlinswavefunctionisnotadequatetofractional
the charge [21]. In GellManns models the fractional
charge of the fundamental particles is the same as in
theexperimentaldata.InLaughlinstheorythereisan
effort to fractionalize the charge from Coulomb
correlations whereas in our theory the fractions
emerge from the spin and the orbital angular
momentum. Jiang et al [31] suggest a Dirac pointin
graphene.Wefindthatallfeaturesinthegraphenecan
be explained on the basis of Schroedinger mechanics
andtheDiracpointsdonotoccurintherealmaerial.

Shrivastava,K.N.,AIPConf.Proc.1250(2010)261264.
Shrivastava, K.N., Introduction to quantum Hall effect,
ISBN1590334191,NovaSci.NewYork(2002).
Shrivastava, K.N., Quantum Hall Effect: Expressions,
NovaSciencePublishers,NewYork,2005,ISBN159454
3992.
Shrivastava,K.N.,AIPConf.Proc.1325,1619(2010).
Shrivastava,K.N.,Proc.Conf.HonourMurrayGellManns
80thBirthday,QuantumMechanics,ElementaryParticles,
QuantumCosmologyandComplexity,ISBN13978981
4335607, ISBN 109814335606, World Scientific Pub.

REFERENCES

Co,PteLtd,Singapore,(2011).pp511517.
Shrivastava,K.N.,Int.J.Mod.Phys.B25,13011357(2011).

Ali, M. M. A. And Shrivastava, K. N. AIP Conf. Proc. 1482,

Shrivastava, K. N., International Journal of Nanoscience

4346 (2012). [International Conference on Fundamental

(WorldSci.)10[IssueNo.3]507514(2011).

and Applied Sciences, 2012, edited by B. AriWahjoedi,

Shrivastava, K. N., AIP Conf. Proc. 1482, 335339 (2012),

R.RazaliandM.Narahari,ISBN9780735410947].

[International Conference on Fundamental and Applied

Essler,F.H.L.,Frahm,H.,Gohmann,F.,KlumperA.and

Sciences, 2012, edited by B. AriWahjoedi, R.Razali and

Korepin, V.E., The Onedimensional Hubbard model,

M.Narahari,ISBN9780735410947].

CambridgeUniversityPress,2005.

19

www.seipub.org/rap

Review of Applied Physics (RAP) Volume 1 Issue 1, December 2012

HewasProfessorintheUniversityofHyderabadfrom1978
till2005.HeservedasProfessorintheUniversityofMalaya
from2005till2011.HeworkedinU.S.A.for5+1/2yearsand
inUK,Japan,theNetherlands,SwedenandSwitzerlandfor
varying lengths of times. He has published 230 papers
withinthelast45years.Heistheauthorofseveralbooks.(i)
Superconductivity:Elementary topics, World Scientific Pub.
Co.Singapore2000,(ii)IntroductiontoquantumHalleffect,
Nova Science Publisher, New York 2002 and (iii) Quantum
Halleffect:Expressions,NovaSciencePublishers,NewYork
2005.

Tsui,D.C.,StormerH.L.andGossard,A.C.,Phys.Rev.Lett.
48(1982)1559.
Thomas,L.H.,Nature117(1926)514.
VonKlitzing,K.,Dorda,G.,andPepper,M.,Phys.Rev.Lett.
45(1980)494.
Zabidi, N. A., Kassim, H. A. and Shrivastava, K. N., AIP
Conf.Proc.Vol.1017(2008),p.326.

Keshav Shrivastava obtained Ph.D. degree from the Indian

Institute of Technology in 1967 and D.Sc. from Calcutta
Universityin1980.

Prof. Shrivastava is a Fellow of the Institute of Physics

London,FellowoftheNationalAcademyofSciences(India)
andmemberoftheAmericanPhysicalSociety.

20