Chapte2

r
Model problemsthat form important
starting points

T h e m o d e lp r o b l e m sd i s c u s s e di n t h i sc h a p t e fr o r m t h e b a s i sf o r c h e m i s t s '
understanding o f t h e e l e c t r o n i cs t a t e so f a t o m s ,m o l e c u l e sc, l u s t e r sa, n o
s o l i d sa s w e l l a s t h e r o t a t i o n aal n d v i b r a t i o n am
l o t i o n so f m o l e c u l e s .

2.1 Freeelectronmodel of polyenes

T h e p a r t i c l e - i n - a - b opxr o b l e m p r o v i d e sa n i m p o r t a n tm o d e l f o r s e v e r a r
relevantchemical situations.

The "particle-in-a-box" modelfor motionin tri'odimensions discussed earlier

can obviouslybe extendedto three dimensionsor to one. For two and three
dilnensions. it providesa crudebut usefulpicturefor electronicstateson surfaces
or in metalliccr1'stals,respectivelv.I sayn.retallic
crystalsbecauseit is in such
systems thattheoutemostvalenceelectrons arereasonably rvelltreatedasrnoyilg
freely.Freemotionu'ithina spherical volumegivesriseto eigenfuncrions thatare
usedin nuclearphysicsto describethemotionsof neutronsandprotonsin nuclei.
In the so-calledshellrrodel ofnuclei, the neutronsand protonsfill separates.
p. d, etc. orbitalswith eachtirpeof nucleonforcedto obey the pauli principle
(i.e.,to haveno more than two nucleonsin eachorbital becauseprotonsand
neutronsare fennions).To remindyou, I displayin Fig.2.l the angularshapes
that characterizes. p" and d orbitals.
This samesphericalbox model has also beenusedto describethe orbitals
of valenceelectronsin clustersof rnetalatomssuchas Cs,,,Cu,,.Na,, and their
positiveand negativeions.Becauseof the metallicnatureof thesespecies. their
r"alenceelectronsare essentiallyfree to roam overthe entiresphericalvolurneof
the cluster.which rendersthis simplernodelrathereffective.In this rnodel.one
thinksof eachelectronbeingfreeto roamu,ithina sphereof radiusR (i.e.,having
a potential that is uniform rvithin the sphereand infinite outsiclethe sphere).
Finally.as noted above,this samesphericalbox model forms rhe basisof the
so-calledshell model of nuclearstructure.In this model.one assumesthat the
protonsand neutronsthat makeup a nucleus,both of u,hichare fermions.occuDv

46
 F r e ee l e c t r o nm o d e l o f p o l y e n e s

_ k - -m l -,,_.-
-k;3-
-n,,,.,J--=]#1'

@
#
W d olLritli
shapes of s, p, and d
f un c t i o n s .
T h ea n g u l a r

sphericrl-boxorbitals(one set of orbitalsfbr proto's.

anothersr"-t for nc.urrons
because thevare distinguishabrc rro' one another).By pracingthe protonsand
neutronsinto thesc- orbitais.t\\,oto an orbital,one achic.n,es a descriptionof the
energyle'els of the nucleus.Excitedstatesareacliie'ed
by prornotinga neutron
or proton fiorn an clccupiedorbital to a lirtuar (i.e.. pre'iously
unoccLrpied)
o r b i t a i .l n s u c ha r n o d c re. s p e c i a i l sv t a b r en u c l e i
a r e a c h i e ' e dw h e n . . c r o s e d _
s h e l l "c o n l i g u r a t i o nssu c ha s l s : o r l s : 2 s r 2 p 6a r c .
r c a l i z e d( e . g . .a H e h a sb o t h
neutronsandprotonsin Isr configurations).
Theorbitalsthatsol'e theSchroclinger equationi'side sucha spherical boxare
notthesamein rheirraclial"shapes"asthes.p, d. etc.
orbitalsof atomsbecause. rn
atoms.thereis an additionalradialpotentialv(r): -Ze2
lr presenr. Honever,
theirang'larshapes arethesameasin atomicstructure because. in bothcases. the
potentialis independent of d andry'.As theorbitalprotsshou,naboveindicate.the
angularshapesofs. p. andd orbitalsdisplayvaryingnumbers
ofnodal surthces.
Thes orbitalshavenone.p orbitalshaveone.a'd d orbitals
havetwo.Analogous ro
how the numberof nodesrelatedto the total energy
of the particleconstrained
to the ,r, -uplane,the nurnberof noclesin the angular
wave functionsindicates
the amountof angularor rotationalenergy.Orbitals
of s shapehaveno angular
energy,thoseofp shapehavelessthendo rl orbitals.
etc.
48 M o d e lp r o b l e m st h a t f o r m i m p o r t a n ts t a r t i n gp o i n t s

The :z
a t o m i co r b i t a l so f a
c o n . j u g a t ecdh a i no f n i n e
c a r b o na t o m s .

one-dinrensional freeparticlemotionprovidesa qualitatircll,correcrpicturc

fbr :r-electronmotion along tlie p" orbitalsof delocalizedpolvenes.The one
Cartesian dir.nension thencorresponds to rnotionalongthe delocalized chain.In
such a rnodel.the box length z is relatedto the carbon-carbonboncllength
R and the number N of carboncentersin'olved in the delocalizednetrvork
I : (A - l)R. In Fig. 2.2, sucha conjugatednetu'orkinvolvingnine centers
is depicted.In this exan.rple.the box lcngthrvouldbe eighttin.resthe c-C bond
length.The ei-eenstates ry',,(l)andtheirenergies8,, represent
orbitalsinto u,hich
electrons areplaced.in the examplecase.ifnine z electrons arc present(e.g..as
in the 1.3.5.7-nonatetraene radical),the groundelectronicstatewould be reprc-
sentedby a total vnavefunctionconsistingof a productin lltich the lorvestfour
ry'sare doubly occupiedand the fifth ry'is singly occupied:

tU: rlpt[y pltzar[lptJt3atltsBt.aullct']1tr:a. (l.l )

The ---component angularmomentumstatesof the electronsare labeledq andt')

as discussed earlier.
A productlr,avefunctionis appropriatebecausethetotal Harniltonianinvolves
the kineticplus potentialenergiesofnine electrons. To the extentthat this total
ener-gycan be representedas the sum of nine separateenergies,one for each
electron,the Hamiltonianalloll'sa separationof r,'ariables

H=LHU) e.2)

in which eachH(j) describesthe kinetic and potentialenergyof an individual

electron.Recall that when a partial differentialequationhas no operatorsthat
couple its different independentvariables(i.e., when it is separable).one can
use separationof variablesmethodsto deconposeits solutionsinto products.
Tlrus,the (approximate)additivity ofllimplies that solutionsof HtU : EV are
productsof solutionsto

H ( . i ) l / ( rj ) : E i t t i ) . (2.3)

The two lowestn-excited stateswould correspondto statesof the form

tlt^ = ltptlrlplrzqrht')r\grlrplraug5plt5u, (.2.4)

and V'* = tltlutlrlpllrglr:frltza*tlltqctt!+fllttu, (2.5)

 F r e ee l e c t r o nm o d e l o f p o r y e n e s
49

wherethe spin-orbitals(orbitalsmultipliedby a or B) appearingin the above

productsdependon the coordinatesofthe variouselectrons.For example,

Irstlrt frtlzatlrzfrlr:ctrL:f lr+ulrsfrlrsa (2.6)

denotes

I'p(r t)l' | fl((2')l/2a(r3)l/zp,' t)rt4qG)1h

fl(r)tltaa(r1)!t5B(rs)ry'5a(re
). (2.j)
The electronic excitation energies from the ground state to each of the
above
excited stateswithin this model would be

. r; T<2
) ,t I J 421
I
2 n tl L 2 Lt I
t : f a": <:"1
and AE'':n:--l -'I ( 2 . 8)
2n lL: Lt J
It turnsout thatthis simplemoderof z-electronenergies providesa quaritatively
correctpictureof suchexcitationenergies. Its simplicityallowsone,for example.
to easilysu-sgest how a molecule'scolor (as reflectedin the complemenrary
color of the light the moleculeabsorbs)variesas the conjugationlengthr.
of
the moleculevaries.That is, longerconjugatedmoleculeshave lower-ener_ey
orbitalsbecauser: appearsin the denominatorof the energyexpression.
As a
result,longerconjugatedmoleculesabsorblight of lowerenergythando
shorter
molecules.
This simpleparticle-in-a-box modeldoesnot yield orbitalenergiesthatrelate
t o i o n i z a t i oenn e r s . i eLsr n l e 'rshsep o r u ' n t i "ai ln s i d et h eb o x " i s s p e c i f i e d .
choosin_s
thevalueof thispotentialll suchrhat ,i) +
lir2ll2 12n1152 /Clis equalto mrnus
thelow'est ionizationenergyof the 1,3,-5,7-nonatetraene radical,givesenergy1e.,,_
els(asE : I/o* lt2h2 l2ntl[n:7tr1; whictrcanthenbe usedasapproximations
t o i o n i z a t i t lenn e r S i e s .
The individual;r-molecularorbitals
JtJ
r = (i)'",in(f ) ( t . e' ), , G [ 3
are depictedin Fig. 2.3 for a model of the 1.3.5-hexatriene
for u'hich "bor
z-orbital sysrem OtJ
the length"I is fivetimesthe distanceR6. bet,"veen neighboring
parrsof carbonatorns.Themagnitude of theAthc-atom centeredatomicorbitalin
the nth:r-molecularorbitalis
-eivenby (2lL)1/2 sin(nrkRlc/ I). In this figure,
positiveamplirudeis denotedby the clear spheres.and
negativeamplitudeis
shownby the darkcnedspheres.where two spheresof rike shading
n='GSgEgD
overrap,the
wave tunction has enhancedamplitude:lvheretwo spheresof different (2iL)t 2 sin\nt.rL); Z :,i x Rc.c
shading
overlap.a nodeoccurs.once again.we notethatthenumberofnodes
increases as Thephases
one rangesfiom the lowestenergyorbitalto higherenergyorbitals.
The readeris of the six molecular
onceagainencouraged to keepin mind this ubiquitouscharacteristic
of quanrum orbitalsof a chain
mechanicalu,avefunctions. containingsix atoms
50 M o d e l p r o b l e m st h a t f o r m i m p o r t a n ts t a r t i n gp o i n t s

{
e

Notle

The nodal
patternfor a chain
c o n t a i n i n gs e v e na t o m s .

This simplemodelallor.l,s
oneto estimatespindensitiesat eachcarboncenrer
andprovidesinsightinto which centersshouldbe rrost amenable to electrophilic
or nucleophilicattack.For example,radicalattackat the cs carbonof thc nrne_
atom nonatetraene systemdescribedearlierwould be more lacile for the ground
statev thanlor eitherv* or w'*. In the former.the unpairedspindensityresides
in ry'5.u'hichhasnon-zeroamplitudeat the Cs site,r-: L
12.ln V* and V,*, the
unpaireddensityis in y'raandr,lr6.
respectively.
both of which ha'e zerodensityat
C5.Tlresedensitiesreffectthe values(2/L)1"2si'(rzARcq/Z of theamplitudes
)
for thiscasein which L : 8 x Rcc for, : 5, 4. and 6, respectivell,. plotsof the
wavefunctionsfor n rangingfrom l to 7 are shor,,,n in anotherforrnatin Fig. 2.4
vyherethe nodal pattern is enrphasized.I hope that by now the studentrs not
temptedto askhow theelectron"gets" from oneregionofhigh amplitude,throu-eh
a node,to anotherhigh-amplituderegion.Remernbersuchquestionsare casr
in
classicalNewtonianlanguageand arenot appropriatewhen addressingthe u,ave-
like propertiesof quantummechanics.

2.2 Bandsof orbitalsin solids

Not only doestheparticle-in-a-boxmodeloffer a usefulconceptualrepresentatlon
of electronsrnoving in polyenes,but it arsois the zeroth-order
model of band
structuresin solids.Let usconsidera simpleone-dimensional.,crystal"
consisting
 B a n d so f o r b i t a l si n s o l i d s

\ .l
r
'-q*C7

^a
' b ' t ' 6 ' h ' b ' ' } : t " s } } " b ' \ \ Theenergy
l e v e l sa r i s i n gf r o m 1 , 2 ,
S,-$S-*i*$l,-gE*'ed*r*Si-*Sil;'@ r-*lsi*-*qe$,- 3, 4,and an infinite
n u m b e ro f o r b i t a l s .

of a largenumberof atontsor ntolecules, eachr,r,itha singleorbital(the spheres

shorvn)thatit contlibutesto thebonding.Let us arran-ue thesebuildin_r: blocksin
a regular"lattrce"assho','"n in Fig.2.5.ln thetop fbur rorvsof thisfigurewe show
thecasew'ithl. 2. 3. and-l bLrilding blocks.To the lefi of eachro\v.wc displavthe
ener-s)- splittingpatterninto which the buildingblocks'orbitalsevolr,eas rhey
overlapandfbrrndelocalized rnolecularorbitais.Not surprisingly. for r : 2, one
findsa bondingandan antibondingorbital.Forn : 3. one hasoneboncring. one
non-bondin-9. andoneantibonding orbital.Finally.in the bottomro\\i.wc attempt
to show "vhathappensfor an infinitelylong chain.The kev point is that the
discretenumberof molecularorbitalsappearingin the l,.l orbitalcasesevolves
rntoa continuumof orbitalscalleda band.This bandof orbitalenergiesranses
from its bottorn(rvhoseorbitalconsistsof a fullll in-phasebondin-e combination
of the buildingblock orbitals)to its top (r.vhose orbital is a fully out-of-phase
antibondingcornbination). In Fig.2.6rveillustratethesefuily bondingandfully
antibondingbandorbitalsfbr two cases- the bottominvolvings-typebuilding
blockorbitals.andtliu'topinvolvingp-typeorbitals.Noticethatwhentheenergy
gap betweenthe buildingblock s and p orbitalsis largerthan is rhe dispersron
(spread)in energyrvithinthe bandofs or bandofp orbitals,a bandgap occurs
betweenthe hishestrnemberof the s bandand the lowestmemberof the n band.
The splittingbetr,veen the s and p orbitalsis a propertvof the individualatoms
comprisingthe solid and variesamongthe elementsof the periodictable.The
dispersionin energiesthat a given bandof orbitalsis split into as theseatoniic
3l M o d e l p r o b l e m st h a t f o r m i m p o r t a n ts t a r t i n gp o i n t s

Fullranlibonding

f Io^-''-
I Sand [aD '
l-trllrh"lrdirr'

t@*,w,*.,@".
I in .ntii F r r l l rr r n r i h o l r d r r r r :
I

lge ffi ffi ffi ffi

6

s band
@
Thebonding
t h r o u g ha n t i b o n d i n g
e n e r g i e sa n d b a n d
o r b i t a l sa r i s i n gf r o m s
andfrom p orbitals.
F u l J vb o n d i n g

orbitalscombineto form a band is determinedby hou,srronglvthe orbitalson

neighboringatomsoverlap.Small overlapproducessmall dispersion. and larse
overlapyieldsa broadband.
Depending on horvmany'alenceelectrons eachbuirdingblockcontributes, the
variousbandsformedby overiappingthebuildingblock orbitalsof theciinstrtuent
atomswill be filledto variouslevels.Forerample,if eachorbitalshou,nabovehas
a singlevalenceelectronin an s orbital(e.g..as in the caseof the alkalimetals).
the s-bandu'ill be half filled in the ground stateu,ith u and paired electrons.
B
Suchsystemsproducevery -{oodcondumorsbecausetheir partially filled bands
allor.l'electronsto move with very little (e.g.. only thermal) excitationamong
other orbitals in this sameband. on the other hand,for alkaline earth sysrems
rvithtwo s electrons per atom,the s bandwill be conrpletely filled.in suchcases.
conductionrequiresexcitationto thelowestmembersofthe nearbyp-orbitalband.
Finally,if eachbuildingblock werean Al (3s23pry atom,the s bandwtruldbe
full andthe p bandwould be half-filled.Sysremswhosehighesterlergyoccupied
band is completelyfilled and for which the gap in energyto the lor.r'est unfilled
bandis largearecalledinsulatorsbecause theyha'e no way to easily(i.e..with
Iittle energyrequirement)promote some of their higher energv electronsfrom
orbitai to orbital and thus effect conduction.lf the band gap betrveena filled
bandand an unfilled bandis small,it n.raybe possibrefor therr.nalexcitation(i.e..
collisionsrvithneighboringatomsor molecules) to causeexcitationof electrons
 D e n s i t i e so f s t a t e si n o n e ,t w o , a n d t h r e ed i m e n s i o n s
53

The valence
a n d c o n d u c t i o nb a n d s
and the band gap.

from the formerto the lattertherebvinducingconductivebehavior.An erample

of sucha caseis illustrared
in Fig.2.7.In contrast.svstemswhosehi-shest ener,sy
occupiedbandis partiallvfilled areconductorsbecausethey havelittle spacing
amongtheiroccupiedand unoccnpied orbitals.
To form a senriconductor.onestarts',vithan insulatorasshownin Fig.2.gwith
its filled(dark)bandanda bandgapbetrveen thisbanclandits empty(clear.) upper
band.lf this insulatorrnaterialweresynthesizecl r,vitha smallamountof ,.dopant"
whose"'alenceorbitalshaveener-uies betu,een the fillecland emptybancisof the
tnsulator. onemavgenerate a semiconductor. If thedopantspecieshasno valence
electrons (i.e..hasan emptyvalenceorbitaly,it givesriseto an emptybandlying
betweenthe filledandemptybandsof the insulatoras shownin Fig.2.ga.In this
case.thedopantbandcanactasan electronacceptorfor electrons ercited(either
thermallvor by light) fiom the filled bandinro rhe dopantband.once ele*rons
enter the dopant band chargecan flow and the systembecomesa conductor.
Anothercaseis illustratedin Fig. 2.8b.Here,the dopanthasits own bandfilled
but lies closeto the emptybandof the insulator.Henceexcitationof electrons
from the dopantbandto the empty bandcan inducecurrentto flow.

2.3 Densitiesof statesin one, two, and three dimensions

W h e n a l a r g e n u m b e r o f n e i g h b o r i n go r b i t a l s o v e r l a p , b a n d s a r e
formed' However,the nature of these bands is very different in different
dimensions.
54 M o d e l p r o b l e m st h a t f o r m i m p o r t a n ts t a r t i n gp o i n t s

,-
!

The filled
and empty bands,the
band gap, and empty
acceptoror filled donor
oanos.

Beforeleavingour discussion ofbandsoforbitalsandorbitalenergiesin solids.

I wantto address the issueofthe densityofelectronicstatesandthe issueofrvhat
deternrines the energyrangeinto which orbitalsof a givenbandu,ill split.First.
lett recall the energyexpressionfor the one- and tr','o-dimensional
electronin a
box case,and let'sgeneralize it to threedimensions.The generalresultis

E:Lnlr2ti1(2mLj). (2.10)

where the sum over./ runs over the nurnberof dimensions(one,two. or three),
and L.1 is the length of the box along the 7th direction. For one dimensron,
one observesa patternof energylevelsthat grows with increasingn, and whose
spacingbetweenneighboringenergylevelsalsogrows.However,in tr,,r,o andthree
dimensions,the patternof energylevel spacingdisplaysa qualitativelydifferent
characterat high quantumnumber.
caseand,for simplicity,let'susea "box"
consider first the three-dimensional
that has equallengthsidesZ. In this case,the toral energyE is th2#l2nrL:1
times Qrl + nf, + n).fne latter quantitycan be thoughtof as the squareof the
lengthof a vectorR havingthreecomponents,?_r, nt,, ft:. Now think of three
cartesianaxeslabelednx,nt,,andn, andview a sphereof radiusR in this space.
The volume of the I /8 spherehavingpositivevaluesof ir.., r ,,,andn_-and having
 D e n s i t i e so f s t a t e si n o n e ,t w o , a n d t h r e e d i m e n s i o n s 55

radiusR is ll8(4/3;rRr).Because eachcubehavingunit lengthalongtheilr, t'rr.,

and il, a-rescorrespondsto a singlequantumwave function and its energy.the
toralnumberl/,"'(6) of quantumstateswith positivefrx, n,., and n. and with
energybetweenzeroand E : (h2n2l2m L21R2is

:;(i"l#1")
i(]'-') (2",
The numberof quantumstateswith energiesbetr.r.een E and E + dE is
whichis thedensityQ(E) of states
(dilibrldE)dE, nearenergyE:
tl+ [tart:1"]r, ,).
o ,'-' : (t.12)
8\ ; " L ; = l a ' )
Noticethat this statedensityincreases as E increases. This meansthat,in the
three-dimensional case,the numberof quantumstatesper unit energygro$s;
in other words,the spacin-ubetweenneighboringstateenergiesdecreases, very
unlike the one-dimensional case rvhere the spacingbetweenneighboringstates
growsasn andthusE grows.Thisgrowthin statedensityin thethree-dimensional
caseis a resultofthe degeneracies andnear-degeneracies thatoccur.Forexarnple,
t h es t a t e w
s i t h n , , n . t .n. : : 2 . l . I a n d l . 1 , 2 , a n d 1 , 2 , I a r ed e g e n e r a taen.d
t h o sw e i t hn r . n \ .n. - : 5 , 3 . I o r 5 . 1 . 3o r 1 , 3 , 5o r 1 . 5 . 3o r 3 . 1 . 5o r 3 . 5 ,I a r e
degenerate andnearlydesenerate to thosehavingquantumnumbers4. ,1.I or l.
;1.:1.or 4. l. .1.
In the two-dimensional case.degereracies also occur and causethe density
ofstatesto possess an interesting6 dependence. In this case.we think ofstates
havingenergyE : (ti2ir2 l2m L21R:,butrvith Rl : rri + ni. The total number
ofstateshavinscnersvbetweenzeroand 6 is
, t ]nL: t
. \ . . . .: , rl r n - : t - t (2.ri)
\,r; )
E and E -t d E is
So, the densitl,of statesbeti.veen

e r E r+: + : + '( 1 r ' 1

LtL \ lf1- ,/
(2.14)

That is. in this two-dimensionalcase.the numberof statesper unit energyis

constantfbr high E valLres
('uvherethe analysisaboveappliesbest).
This kind of analysisfbr the one-dimensional casegives

( 2 l.s )

so the statedensitybetrveenE and E -t d E is

, . t'l
e r t ) : r r : . ( - J ! !)- u - ' , , . (2.16)
\ tr:T- /
which clearlyshowsthe lvideningspacing.and thus lowerdensity,as one goes
to higherenergies.
56 M o d e lp r o b l e m st h a t f o r m i m p o r t a n ts t a r t i n gp o i n t s

Two ls
orbitals combine to
produce a o bonding
and a o" antibonding
m o l e c u l a ro r b i t a l .

Thesefindingsaboutdensitiesof statesin one.trvo.andthreedimensionsare

irnportantbecause.in variousproblemsone encountersin studyingelectlonic
statesof extendedsvstemssuchas solidsand surfaces. one needsto knou,how
the numberof statesavailableat a siven total energy.Evariesll ith 6. clearlv. the
ans\ver to thisquestiondependsuponthedimensionality of the problem,andthis
fact is rvhatI lvantthe studentsreadingthis text to keepin mind.

2.4 The most elementarymodel of orbital energysplittings:

Hiickelor tight-bindingtheory
Nou'let'serarninewhatdeterrnines theenergyrangeinto which or-bitals (e.g, p"
orbitalsin polyenesor metals or p orbitalsin a solid)split.To begin.considertwo
orbitals.oue on an atom labeledA and anotheron a neighborinsatom labeled
B; theseorbitals could be. for example,the 1s orbitals of two hydrogenatoms.
such as Fig. 2.9 illustrates.However,the two orbitals could insteadbe til'o p"
orbitalson neighboringcarbonatomssuch as are showrrin Fig.2.l0 as they
form bondingand n* antibondingorbitals.In both of thesecases,we think of
formingthe molecularorbitals(Mos) @6as linearcombinations of the aronric
orbitals(Aos) x, on the constituentatoms,and we expressthis mathematicallv
as follou's:
, s-^
@^: ) Lr'.,Y,. (2.t7\

wherethe C1.,,orecalledlinearcombinationof atomicorbitalsto forn molecular

orbital (LCAo-Mo) coefficients.The Mos are supposedro be solutionsto the
Two atomic Schr<idinger equationin which the HamiltonianH involvesthe kinetic ener_ey of
p- orbitalsform a
the electronas well as the potentialsZ1 and zq detailingits attractionto the left
b o n d i n gz a n d
a n t i b o n d i n gz * and right atomic centers:
m o l e c u l a or r b i t a l .
H=-lil2ny2-rL*lq ( 2 .I 8 )
 T h e m o s t e l e m e n t a r ym o d e l o f o r b i t a le n e r g ys p l i t t i n g s
57

In contrast,theAos centeredon the left atomA aresupposedto be solutionsofthe

Schrodinger equationwhoseHamiltonianis H : -/1?l2nty2 -f Vt,andtheAOs
onthe right atornB haveH : -h2 l2mV2 + Vn.Substitr.rting 0* :L.Cx.uXu
into the MO s Schrodinger
equationHQr : €rQr and then multiplyingon the
leftbythecomplexconjugateof x7,andintegratingoverther,0 andQcoordinates
of the electronproduces

(2.19)

Recall that the Dirac notation (a I b) denotesthe integral ofa* and 6, and (a loplb)
denotesthe integral ofa* and the operatorop acting on D.
In what is kno'uvnas the Hiickel model in organic chemistry or rhe tight-bin<1ing
model in solid-state theor1,,one approximates the integrals entering into the above
set of linear eqr-rations
as follows:

(i) The diagonalintesralt,X,,1-tt)

l2nY) + Li, + tr11X6)involvingthe AO centered
on thc right atomand labeled17,is assumed
to be equivalentto
-h! l2nV) * IrnlXr.),
(,X1, which meansthatnet attractlonoi-thisorbitalto the left
atomiccenteris neglected. Moreover.this integralrs approximated in termsof the
bindingenergv(denotedcr,not to be confuscdrviththe electronspinlirnctrona)
for an electrttnthat occupiesthe 17,orbital:|,y,,-li l2my) ,t l/plXr) = trr,.The
physicalrreaningof ar is thc kineticenergvof the elcctronin 17,plusthe
attractionof this elc-ctrorr to the'rightatomiccenter*,hileit residesin 17,.Of
coursc.an analoqous approximation is madefor the diagonalintegralinvoliing
x , , ; Q , , ) - / t : l 2 r n Y : + L ' r . 1 x ,-, ya , , .
(ii) The off-diagonalinregrals\xt,) - F l2nv. + ti + L\11x,,\1 areexpresserJ in
termsof a pararrerer rvhichrelatesto the kineticand potentialenereyof the
fJ,,.7,
elcctronu,hilcit reside-sin the "overlaprcgion"in which both X,,and X7,are
non-vanishing. This regionis shorvnpictoriailyin Fig.2. I 0 as the regionrvherethe
left and rieht orbitalstouchor overlap.The magnitudeof 13is assumed to be
proportionalto the ovcrlapS,,.7, betrveenthe two AOs: .t,.r,: e, I Xr,).lt turnsout
that f is usuallva negatir,,e
qr.rantity.
which can be seenby writing it as
(/.t,i-lt:l2ntv: + I\<1il i (,xr)t'tlx). Sincex,, is an eigenfuncrionof
-il 12ntV: a lzohavingthe eigenvaluea,,, the first term is equalto s, (a negauve
quantitv)tinres(X1,I X,,).the overlapS. The sccondquantity\Xr,ll/rlX")is equalto
the integfalof the overlapdensity17,Q.)tr"(r) multipliedby the (negative)
Coulorrrb potentialfbr attractir"e interaction of the electronwith the left atomtc
center.So.r,r'henever 17,(r.)and 1,,(r)havepositiveoverlap,p r.r'illturn out
negattVe.
(iii) Finally,in the mostelernentary I-ltickelor tight-bindingmodel,the overlap
integralst,X,,I ft,l - t,./, are neglectedand set equalto zero on the right sideof
the matrir ciqenvalueequation.However,in some HLickelmodels,overlap
betweenneiehboringorbitalsis explicrlvtreatc'dso in someof the discr.rssion
below '"veu'ill retain .!, ,,.
58 M o d e l p r o b l e m st h a t f o r m i m p o r t a n ts t a r t i n gp o i n t s

With theseHiickel approximations. the set of equationsthat determinethe

orbitalenergies €6 andthe corresponding LCAO-MO coefficients are\\'rit-
C'(.(?
ten for the trvo-orbital
caseat hand as in the first 2 x 2 matrix cquationssho\\'n
below:

[l f "i "' J] L[ c:l, l: ,tI. sl :r Jl[L:( I, { J (2 . 1 0

)

u,rittenas
*'hich is sor.netimes

| " - ' / r- . s .[1. . . ]- Ll nol . l ( 2 . 2I )

[ r - . s " - . . ]L . - . 1
of zerooverlapto
reduceu'ith the assulxption
Theseequations

[ " , . l [ , -=l . I ror l Ll I. 'ci l, ' ' l (2 . : 2)

Lf,"_L
l c*.1 L0
Thea parameters areidenticalif thetwo AOs x,, andXr,areidcntical.asr"'ould
be the casefor bonding betrveenthe tu'o ls orbitalsof lwo H atolltsor nt'o )11-
orbitalsof tu'o C atomsor tr.vo3s orbitalsof tr",oNa atoms.If the left and right
orbitalswerenot identical(e.g.,for bondingin Hel-l+or for the:r bondingirr a
C-O group).their o valueswould be differentand the Hiickel matrix problen.t
lr'ouldlook like:

[ " n l [ ' l : , I sr . ]l [ c I r l l i t

lf,"lLCo.l Ls LCRI
To find the MO energiesthat resultfrorn combiningthe AOs. one must find
the valuesof e for which the aboveequationsare valid. Takin-qthe 2 x 2 matrix
consistingof e times the overlapmatrix to the left-handside.the aboveset of
equationsreducesto Eq. (2.21). It is knou,nfrom trratrixal-uebra that sucli a set
oflinearhomogeneous equations (i.e.,havingzeroeson theright-handsides)can
havenon-trivialsolutions(i.e.,valuesof C that are not simplyzero)only if the
determinantof the n.ratrixon the left sidevanishes.Settingthis deteminantequal
to zero giles a quadraticequationin rvhichthe e valuesarethe unknor.vns:

(a-e):-(B-oS)r=0. (2.24)

This quadraticequationcan be factoredinto a product

( a - f - e * s S ) ( a+ P - € - s S ) : 0 . ( 2 . 2 )5

which hastrvo solutions

e:(a*p)/(l +S). and e = (a- F)l(l- S). (2.26)

As discussedearlieqit rums out that the B valuesare usuallynegative.so the

lor.restenergysuchsolutionis the e : (a * P)l0 + S ) solution,which givesthe
 T h e m o s t e l e m e n t a r ym o d e l o f o r b i t a le n e r g ys p l i t t i n g s
qq

energyof the bondingMo. Noticethat the energiesof the bondingand antibond_

ing iVlos are not symmetricallydisplacedfrom the value a within this \.erslon
of the Htickelmodelthat retainsorbitaloverlap.In fact,the bondingorbitallies
lessthanp belowcv,and the antibondingMo lies morethanB abovea because
of the I * s and I - s factorsin the respectivedenominators.This asymmer-
ric lou.eringand raising of the Mos relativeto the energiesof the constituent
Aos is commonlyobservedin chemicalbonds;that is, the antibondingorbital
is more antibondingthan the bonding orbital is bon<iing.This is ar.rother im-
portantthing to keep in mind becauseits effectspervadechemicarbondins and
spectroscop)'.
Havingnotedtheeffectof inclusionofAo overlapeffectsin the Hiickel moder,
I shouldadmit that it is far more commonto utilize the simplifieciversionof the
Hiickelmodelin which the s factorsareignored.In so doing,oneobtainspatterns
of N'Ioorbitalenergiesthat do not reflectthe asymmetricsplittingin bonding
and
antibondingorbitalsnotedabove.However.this simprifiedapproachis easier
to
useand ofl'ersqualitativelycorrectMo energyorderin-es. So, let'sproceedwith
our discussion of the Hiickelmodelin its simplifiedversion.
To obtainthe LCAo-Mo coerfcientscorresponding to the bondingandanti-
bondingMos. onesubstitutes thecorresponding a valuesintothelinearequarrons

f";'.1,]tr]:t:]
andsoh'esfor the C,,coefficients(actually,onecan solvefor alr but one c',,.and
t1)1\

thenusenorrnalizationof the Mo to deterrninethefinal c,, ). Forexample.for the

bondingN,lO.lve substitutee : a * B into theabovematrixequationandobtain
two equationsfbr C1 and Cp:

-fCt-t
dCn:0. ( 2 . 2)8
IJCL-pcR-0. 12)9)
Thesetr.voequations areclearlynotindependent;
eitheronecanbe solvedfor one
C in termsof the otherC to sive

Cr = Cn. (2.-r0)
which means that the bonding MO is

Q:CJxr*xil. (2.31)
Thefinalunknor,vn, C1,is obtainedby notingthat@is supposedto bea norrnarized
function (A I A) : l. Within this versionof the Hrickel
model, in which the
overlaps is neglectedthe normalizationof leadsto the
@ followine condrrion:

| : @ | d ) = C l ( t , x t lx L )+ ( x R x R :) ) 2 C l .
AN M o d e l p r o b l e m st h a t f o r m i m p o r t a n ts t a r t i n gp o i n t s

with the final resultdependingon assumingthateach1 is itselfalsonormalized.

. e k n o wt h a tC 1 : 1 1 l 2 ) l : . a n dh e n c et h e b o n d i n gM O i s
S o ,f i n a l l y w

O:,ll2)t' er+x*) rl l'li

A c t u a l l yt.h e s o l u t i o no f I : 2 C ' i c o u l da l s oh a v ey i e l d e dC ' r : - ( l / 2 ) 1 t 2 a n d

then vr,ewould have

O: - , l t l 2 ) t t . ( x r* x n ) (2.3.1)

Thesetwo solutionsarenot independent (oneisjust - I tinresthe other).so only

one shouldbe includedin the list of MOs. Horvever. eitherone is just as goodas
the otherbecause. as shownvery earl1,in this text.all ofthe ph1'sical properties
that one computesfrom a wave function dependnot on ry'but on r/r*{r. So. tu,o
u'avefunctionsthat differ from one anotherby an overallsign factor.as u,ehave
here,haveexactlythe sar.nerlr*tb andthus are equivalent.
ln like fashion.we cansubstitute€ : cr - p into the matrix equationandsolve
for the C1 and Cp valuesthat are appropriatefor the antibondingMO. Doing so
glvesus
i)
Er : 1t121 (xr - Xr.) (2.3
5)

or. alternatively.

O " : ( l l 2 l t ' 2 ( X*n X r ) . (2.36)

Again. the fact that either expressionfor @* is acceptableshowsa property of

all solutionsto any Schrodingerequations;any multipleof a solutionis alsoa
soiution.In the aboveexample,the two "answers"for @*differ by a multiplicative
factorof(-1).
Let's try anotherexampleto practiceusing Hiickel or tight-bindingtheory.In
particular,I'd like you to imagine two possiblestructuresfor a clusterof three
Na atoms(i.e.,pretendthat someonecameto you and askedwhat geometryyou
think sucha clusterwould assumein its ground electronicstate).one linear and
onean equilateraltriangle.Further,assumethatthe Na-Na distancesin both such
clustersare equal(i.e.,that the personaskingfor your theoreticalhelp is willing
to assumethat variationsin bond lengthsarenot the crucialfactor in determining
which structureis favored).In Fig.2.ll, I showthe two candidateclustersand
their 3s orbitals.

L i n e a ra n d
e q u i l a t e r at rl i a n g l e
structuresof sodium
Inmer.
 T h e m o s t e l e m e n t a r ym o d e l o f o r b i t a le n e r g ys p l i t t i n g s o l

NumberingthethreeNa atoms'valence3s orbitalsh, X2,and X3,we thenset

up the 3 x 3 Htickel matrix appropriateto the two candidatestructures:

[ " oe po ll (2.31)
l p
Lo P d)
for the linear structure(n.b.,the zeroesarisebecause11 and Xj do not overlap
and thus have no p coupling matrix element).Alternatively,for the triangular
structure.we find

[ " p e 1
l P" P l (2.38)
L Pp " )
as the Hiickelmatrix.Eachof these3 x 3 matriceswill havethreeeigenvalues
thatwe obtainby subtractinge from their dia_eonals
and settingthe determinants
ofthe resultingmatricesto zero.For the linearcase,doing so generates

(a*e)r -2ft(o -e):0, (2.3e)

and for the trian_qlecase it produces

{ o - r ) ' - 3 d : ( o- e ) + 2 d r : 0 (:..10)
The first cubicequationhasthreesolutionsthat give the MO energies:

s:cv*(2)r:f,. F:(r. and r=d-(21tr..U, (:.,11)

for the bonding.non-bondingand antibondingMOs, respectively.

The second
cubicequationalsohasthrecsolutions

t::a*2f. s:a-fl. and e:u-8. (2 . 4 2 )

So.for the linear and triangularstructures,the MO energypatternsare as shou'n

tn rls. l. Il.

- a - Q)1/2p
-c-B

Energy
o r d e r i n g so f m o l e c u l a r
- a + (z)tt2| o r b i t a l so f l i n e a r( l e f t )
a n d t r i a n g u l a r( r i g h t )
-c+2p s o d i u mt r i m e r s .
62 M o d e l p r o b l e m st h a tf o r m i m p o r t a n ts t a r t i n gp o i n t s

Forthe neutralNa: clusteraboutr".hichvou wereaskedyou havethree'u'alence

electronsto distributeamongthe lowestavailableorbitals.In the lincarcase.we
placetr'"'oelectronsintothe lowestorbitalandone into the secondorbital.Doing
s o p r o d u c c s a t h r c e - e l e c t r o n s t a t e wei trhr retrogtluo6lf : l t u + l ' l , ' l ) + o :
3cv* 2 2t i2B. Altematively,for the triangularspecies.we put trvo electronsinto
the lowestMO and one into eitherof the degenerate MOs resultingin a three-
electronstatewith totalenergl'E :3a f 38. Becauseli is a uegativequantitv.
thetotalenergyofthe triangularstructureis lowerthanthatofthe linearstructure
r
s f n c c - i> I l
^ l -1 .

The aboveexampleillustrates how we canuseHtickel"tight-binding theor)to

makequalitativepredictions(e.g..which of trvo"shapes"is likel.vto be of lor"'er
ener'-uy).
Noticethat all one needsto knou'to apply'sucha model to any set of
atomicorbitalsthatoverlapto form MOs is:

(i) theindividual a (whichrelatcto theelectronegativitv

AO energies of theAOs)and
(ii) thedegreeto whichtheAOscouple(theB pararreters
*'hichrelateto AO or,eriaps).

Lett seeif you can do someof this on your own. Using the aboveresults.
uould you expectthecationNai to be linearor triangular? What abouttheanron
Na.,? Next. I u,antyou to substitutethe MO enersiesback into the 3 x 3 matrix
andfind the C1. C2,and C3 coefficients to eachof the threeMOs of
appropriate
the linearand ofthe triangularstructure.Seeifdoing so leadsyou to solutions
that can be depictedas shor,l'n
in Fig.2.l3" and seeif you can placeeachsetof
MOs in the properenergyordering.

@
The
(-a-a-\
\_\v
orbitalso{
l i n e a ra n d t r i a n g u l a r
s o d i u mt r i m e r s( n o t e ,
they are not energy
ordered).
 T h e m o s t e l e m e n t a r ym o d e l o f o r b i t a le n e r g y
splittings
bJ

Ethylene
with four C-H
bonds,one C-C o bond,
and oneC-C z bond.

No'v. I want to showyou hor'vto broadenyour horizons

and usetight-binding
theoryto describeail of the bondsin a more compricated
moleculesuch as
erhvleneshownin Fi-e'2.14. within the moder described
above, eachparr ot'
orbitalsthat "touch" or overlap-9ivesrise to a 2 x 2 matrix.
More correctry,ail
n of the constituentAos form an n x n matrix, but this
matrix is brokenup rnro
2 x 2 blockswhenevereachAo touchesonly one
otherAo. Noticethatthis did
not happenin the trian-uularNa3casewhereeachAo touchedtwo otherAos.
For
the ethylenecase,the valenceAOs consistof (a) fbur
equivalentC sp: orbitals
that are dire*ed towardthe four H atoms.(b) four
H ls orbitals,1c)t*.o c spr
orbitalsdirectedtowardone anotherto fo'n the
c-c o boncl and (d) two c p.,
orbitalsthatwill forn.rthec-c z bond.This totar
of l2 Aos generates six Hticker
matrcesas shor.vn bero*,:we obtainone 2 x 2 matrix for the c-c
o bondoi-the
form

o.o, d.o,.o,-l
I (2..13
)
| 1,n,..0, a,o, J
and one 2 x 2 matrix lor the C-C z bond of rhe
form

[ ",' r)r,,.]
(l -+4)
lfo. w "n. l
Finally.we alsoobtainfour identical2 x 2
n.ratrices
fbr the c-H bonds:
r -
| 0,p, l,n,.n 1
Q'45)
Lp.n,n ", J
The above marricesrvill then produce (i)
four identical C_H bondin_e MOs
h a v i n g e n e r g ie e :s t / 2 { ( a , * r y c ) _
[ ( a H_ a c ) : * 1 p 2 l r , z ] , f i i i i o u . : a . n t . u f
C-H antibondingMOs havingenergiest* :
I/21@u * a6) f [(ns _ oc)r +
4 7 2 l t i 2 | , ( i i i ) o n e b o n d i n gC - C r o r b i t a l
with E:epr + fl, (iv) a parrner
antibondin-eC-C orbital with e* : dp, _
F, &) a C_C o bonding MO with
t : ospr* p ' and(vi) its antibonding partnerwith e* - cy,pr p. tn att of these
expressions.the B parameteris supposed
to be that appropriateto the specific
orbitalsthat overlapas shownin the matrices.
64 M o d e l p r o b l e m st h a t f o r m i m p o r t a n ts t a r t i n gp o t n t s

If you wish to practicethis exerciseof breakinga lar_ue rnoieculedown into

setsof interactingAos. try to see what Hiicker matricesy.ou
obtain and r.,,tat
bondingand antibondingMo energiesyou obtain for the 'alence
orbitalsof
methane s h o w ni n F i g .2 . 1 5 .
Beforeleai'ingthis discussionof rhe Hrickerrti_cht-binding nroder.I needto
stressthat it has its ffar.vs (becauseit is basedon approximations and rnvolves
neglectingcertainterms in the Schrcidinger equation).For example.it predicts
(seeabove)thatethylenehasfour energeticallv identicalc-H bondingMos (and
four degenerate c-H antibondingMos). Hou,ever, this is not what is seenwhen
photoelectron spectraare usedto probetrreenergies
Methane of theseMos. Likeu,ise.it
m o l e c u l ew i t h f o u r C * H suggests thatmethanehasfour equivalent C-H bondingandantibonding orbitals.
o on o s . r,''hichagainis not true.It turnsout that.in eachof thesetwo
caseslethyleneand
nrethane). theexperiments indicatea groupingof four ilcarlr.iso-ener_qetic bond_
ing Mos andlour nearryiso-energetic antibonding Mos. Horve'cr.thereis some
"splitting"
amongtheseclusrersof four Mos. The splittingscan be interpreted,
within the Hrickelmoder.as arisingfrom couplirgsor interactions
anron_c. for
example.onesp2or sp3orbitalon a givenC atomandanother
suchorbitalon the
s a n r ea t o m .S u c hc o u p l i n u cs a u s et h e , v r r H L i c k eml a r r i xt o
n o tb l o c k _ p a r r i r i o r r
lnto groupsofl x 2 sub-matrices because no*.thereexistoff-diagonalp factors
thatcoupleonepairof directedAos to another. when suchcouplingsareincluded
in the analysis.one findsthatthe clustersofN{os expected
to b.,l.g.nerate are
not. but aresplitjustas thephotoelectron datasuesesr.

2.5 Hydrogenicorbitals
The hydrogenicatom problem forms the basis of much
of our thinking
about atomic structure.To sorvethe correspondingschrodinger
equation
requiresseparationof the r, 0, andg variables.

The Schrcidinger equationfor a singleparticleof nass p movi'g in a central

potential(one that dependsonly on the radial coordinate
r) can be written as

_-
li 1 a: a: a:\ .--.-\
I I -l- +
+ l - l , r , t Ir '/l ( r :' -L- -:+ ' r 2 + : : ) t l t: E l t ' e46)
2g \ 3:i2 ail o = t) ' l ' +
or, introducingthe short-handnotationV2:

-lt2
/2mY2p * Illt = Ep (2.47)
This equationis not separablein Cartesiancoordinates(x,,i,,:) becauseof the
way r, 1', and ; appeartogetherin the squareroot. However.
it is separablei'
sphericalcoordinateswhereit hasthe form

-- !, w . [ a / . . , a / \ l * I a t. a/\ t d:t
L; \'-;/J ;,*, ( ' ' " ,* i - _ . * ; # - r r , ' t , t '
- - t i / 2 r n v : g+ v { / : E V
(2 . 4 8 )
 H y d r o g e n i co r b i t a l s
o5

Subtractingv(r)(r fron both sidesof the equationand multipryingby -2{i

thenmoving the derivativeswith respectto r to the right-handside.one-obtalis
I U / a/i \ | A:,IJ
-#2rt
ur)
- Itu 'tlr- lt ( ,' #)
d r Il * *i a # :
-:- sintt*--|
s , nd H|\ #
(2.+e)
Noticethat,exceptfbr ry'itsell the right-handside of this equationis a functron
ofr only; it containsno d or @dependence. Let's call the entireright-handside
F(rlrL to ernphasizethis fact.
To furtherseparatethe d and@dependence, we multiply by sin?g andsubtract
rhed derivativeterms from both sidesto obtain

:F(,")ry' o - sinoj_*(.'"r#)
sinr (2.s0)
#
Norv we have separatedthe @ dependencefrom the 0 andr
dependence.we
now introducethe procedureusedto separatevariablesin differentiar
equations
and assumery' can be written as a function of times a function
@ of r) and0:
,lr : A@)Q(r.0). Dividin_eby Oe, we obtain

t
* # : ( o , s i nt2) e -s i n e G
$
'"r#)) (2.51)

Now all ofthe / dependence is isoratedon the left-handside;the right-handside

contalnsonly r andd dependence.
wheneverone hasisolatedthe entiredependence on onevariabreas we have
doneabovefbr the @dependence. one can easilyseethat the Ieft- and right-hand
sidesof the equationmust equala constant.For the above
exampte.thetert-
handsidecontainsno r or g dependence and the right-handsidecontainsno p
dependence. Becausethe two sidesare equal,they both must actualiycontain
no
r,0, or @dependence; thatis, theyareconstant.
For the aboveexample,we thereforecan set both sides
equarto a so-called
separation constantthat we call -m2.lt will becomeclearshortlywhy
we have
chosento expressthe co'stant in the form of minus
the squareof an integer.
You may recallthat we stucliedthis same equatron
@ earlierand learnedhow the
integerrr arisesvia the boundarycondition that
Q andQ * 2z representidentical
geometries.

2 . 5 . 1T h e O e q u a t i o n
The resulting@ equationreads(the "symbor is usedto
representsecondcreriva-
tive)

o"+t1:o:o (2.52t
This equationshouldbe familiarbecauseit is the
equationthatwe trearedmuch
earlierwhenwe discussedthe:-component of angularmomentum.So.its further
oo M o d e l p r o b l e m st h a t f o r m i m p o r t a n ts t a r t i n gp o i n t s

analysisshouldalsobe familiar.but, for completcness.

I repeatmuch of it. The
aboveequationhasas its mostgeneralsolution

tnlr) ( 2 . 5)3
Q = Ae""Q + Be

Becausetlrervavefunctionsof quantummechanics representprobabilitydensi-

ties.they r.nustbe continuousand single-valued.
The lattelcondition,appliedto
our <Dfunction.meansthat

O(d):Q(2;t*4tl r r 5 4l

. 1 e i " " 1 ' (-l ' J r i n t ,+' B e i * a ( l - c : " " - ; : g. (2.5-5)

This conditionis satisfied only whentheseparation constantis equalto an integer

m : 0 , + 1 . + 2 . . . . a n dp r o v i d e sa n o t h eer x a m p l eo f t h e r u l e t h a l q u a n t i z a t i o n
comesfrom the boundarycor.rditions on the r.vavefunction. llere nr is restricted
to certaindiscretevaluesbecause the wavefunctionmustbe suchtl.ratwhenyou
rotatethrough2:r aboutthe :-axis. you must get back \\4ratyou startedvn,itl.r.

2 . 5 . 2T h e O e o u a t i o n
Norv returningto the equationin which the @dependence was isolatedfrom the
r and0 dependence and rearrangingthe 0 termsto the left-handside,rve have

1 ii / dO\ nr:O
- : Ftr)Q (2.-56)
""p a6 \sin6-i/ *t
In this equationwe haveseparated
0 andr variationssowe canfurtherdecompose
tlrewavefunctionby introducingQ : O(6)R(r), which yields

1 1 at r)G)\ lrl FtrtR

l s i n 6 -l- *: ^ =-i.
O ri"B ap \ do/ sln-H ^

where a secondseparationconstant,-),, has been introducedonce the r and

0 dependentterms have been separatedonto the right- and left-hand sides,
respectively.
We now can write the d equationas

##(''",#)-#--,(,, ( 2 . 5)8

wheren is the integerintroducedearlier.To solvethis equationfor O, we make

z : cosI and P(z) : O(9), so JT -= : sind, and
the substitutions

0 020 . d. -, d
= -sln (2.59l,
Ae 60 0z dz
 H y d r o g e n i co r b i t a l s ot

Theran-qeof valuesfbrd was0 < I < z, so the rangefor: is -l < : < l. The
equationfbr O, whenexpressedin termsof P and:. becomes

I / ,/Pt ,r,]P
- l r l - -r r--- l (2 . 6 0
21"-='E)-t--=t-)'P:o )

Noi.r'we can look for polynomial solutionsfor p, because: is restrictedto be

lessthanunity in magnitude.lf m :0. we first let

p -- f . , - r
/-"t- (2.61)
t=1)

intorhedifferential
andsubstitute equationto obtarn
-
\-rl - 1 ) r l . J - l rl ,t r , - : -I - ' - L
\ - ,. f, L l ,l .k, t r l : "- ; f
/-,"'-',', t t . : a= 0 . (2.62)
{=1,

Equatinglike powersof ; gives

ttt(k(k]_l)-^\
( 1 . 6 i)

Note that for lar_sc-

valuesof l

,,^_,- tr(t+i) _,
,k (2.64)
fr,(l+i)(r+i)-'
Sincethe coefhcicntsdo not decrease with f for larget. this serieswill diverge
for-- : *l unlessit truncates at finiteorder.This truncationonly happensif the
separatron constanti obe-v"s ), : l(l * l). where/ is an integer.So. onceagtin.
we seethata boundarvcondition(i.e..thatthe wavefirnctionnot diverge
andthus
be normalizable in this case)gil'esriseto quantization. In thiscase.the valuesof
i arerestricred to 1(/ * I ): before.we sawthat,r is restricted to 0. * l. +2. . . .
Sincethe'above recursionrelationlinkseveryothercoefflcient. we canchoose
to solvefor the evcnand odd functionsseparately. choosingd1,and thendeter-
miningall of the evenr1rin termsof this,q,. fbllor^,ed by rescalingall of these
ar to lrakc the functionnormalizcd. generates an e\.ensolution.choosinga7and
determiningall of the odd a1 i' like rnannereenerates an odd solution.
For / : 0. the seriestruncatesafter one term anclresultsin po(:) : L For
/ = I t h es a m et h i n ga p p l i e as n d p 1( : ) : ; . F o r1 - 2 . a t : _ ' 6 + : _ 3 a o
, so
o n e o b t a i n sP : : 3 2 ) - l . a n d s o o n . T h e s ep o l y n o m i a l ,s. . . . l l . d
Legenclre
polvnominls.
For the more generalcasewherem I 0. one can proceedas aboveto gener-
atea polynomialsolutionfbr the o function.Doing so resultsin the
followine
solutions:

- ,,, ",.1,,) tp.t -\

t \ - l
I > t ' t t _ r _ t t _ l , i
| | t- t - \ r
- _ ./ (2.(,5)
d:tnt
o.J M o d e lp r o b l e m st h a t f o r m i m p o r t a n ts t a r t i n gp o i n t s

Thesefunctionsarecalledassociated Legendrepolynomials. andtheyconstitute

the solutionsto the O problemfor non-zerorr r.'alues.
TlreaboveP andei''4'functions. whe're-expressed in termsof 0 [email protected]
thefull angularparrofthe wavefunctionfor anycentrosymmetric potential.These
tQ exp(inrQ),and
Q): pi,,rcose)er)
solutionsare usualll,written as Y1.,,,(0.
arecalledsphericalharmonics.Theyprovidethe angularsolutionof the (r. 0. d)
Schrodinger equationfor anl problemin rvhichthepotentialdepends onl1,sn 11-r.
radialcoordinate. Suchsituationsincludeall one-electron atomsand ions ie.g.,
H, He*, Li*+. etc.).the rotationalmotion of a diatomicrnolecule(u,herethe
potentialdepends only on bondIengthr). themotionof a nucleonin a spherically
symmetrical "box" (as
occursin the shellmodelof nuclei).andthe scattering of
two atoms(u'herethe potentialdependsor.rlyon interator.llicdistance).The [.,,,
functionspossessr,aryingnumbersof an-{ularnodes.u,hich.as noted earlier.
give clear signaturesof the angular or rotationalenergy content of the wave
function.Theseangularnodesoriginatein the oscillatorynatureofthe Legendre
and associated Legendrepolynomialsp,"'(cosd);the higher/ is. the more sign
changesoccurwithin the polynomial.

2 . 5 . 3T h e F e q u a t i o n
Let usnowturnour attentionto theradialequation, whichis theonlyplacethatthe
explicitform ofthe potentialappears.Using our earlierresultsfor the equatlon
obeyedby the R(r) functionand specifyingv(r) to be the coulomb potential
appropriatefor an electronin the field of a nucleusof charge*Ze, yields

l d /.dR\ f2u/_ Z e z \ t ( t+ t t l r
Ft\'-,1,)+Lf\";)- .: Jn:o {266)
we cansimplify thingsconsiderablyifwe chooserescaledlengthandenergyunits
becausedoing so removesthe factorsthat dependon p,li,and e. we introducea
new radial coordinatep anda quantityo as follows:

,, 7) .,1
, : (-t:u \i ,-. , ) F Z C

\ r / \ 2 . 6 7)
2Ett
Noticethat if E is negative,asit will be for boundstates(i.e..thosestateswith en-
ergybelowthatof a freeelectroninfinitely far from thenucleusandwith zerokin-
etic energy),p is real.on the otherhand,if E is positive,asir will be for statesthat
lie in the continuum,p will be imaginary.Thesetwo caseswill give riseto quali-
tativelydifferentbehaviorin the solutionsofthe radialequationdevelopedbelow.
We now definea function S suchthat S(p) : R0.) and substituteS for R to
obtain

i j ; G f l . ( - i- ' # + 1 ) s : o (2.68)
 H y d r o g e n i co r b i t a l s 69

The differentialoperatortermscan be recastin severalways using

I ,/ / .,/S\ d2s zds I d1
- (2.69)
;-.T ,,tpS).
,t,tr\tO)= ap- pLtp pLlp-

The strategythat we norv follow is characteristicof solving secondorder

equations.we will examinethe equationfor s at large and small p
diff-erential
values.Havingfound solutionsat theselimits,we will usea powerseriesin p to
"interpolate"bet'"veen thesetwo limits.
Let us begin by examiningthe solutionof the aboveequationat small values
of p to seehow the radial functionsbehaveat smallr. As p + 0. the secondterm
in thebracketsin Eq.(2.68)will dominate.Neglectingthe othertwo termsin the
brackets.we find that,tbr srnallvaluesof p (or r ), the solutionshouldbehavelike
pL andbecausethe tunctio'must be normalizable. we must haver > 0. Since
z canbe any non-negativeinteger.this suggeststhe followin-emoregeneralform
for S(p):

S(p) x pLe-u1 (2.70)

Thisform rvill insurethatthefunctionis normalizable sinceS(p) * 0 asr * eq,

forall l, as long as p is a realquantity.If p is imaginary,sucha form may not
be normalized(seebelowfbr turtherconsequences).
Turningnow to the beha'ior of s for largep. rvemakethe substitutionof .!(p)
intotheaboveequationandkeeponlv thetermswith the largestporverof p (e.g.,
thefirstterm in bracketsin Eq. (2.68)).Upon so doing.rveobtainthe equation

a2pt g un: o ",' t : . 7| )

!o''
whichleadsusto concludethattheexponentin thelarge-pbehaviorof s i. u : j.
Havingfbundthe small-and large-pbehaviors of s(p), we cantakes to havethe
followingform to interpolatebetrveen
largeand smallp-values:

S (p ) : p L e - / ' t 2P 1 p 1 . \/-.tL)

wherethe function P is expandedin an infinite powerseriesin p as p(p) :

Lorpr. Again substitutingthis expressionfor S into the aboveequationwe
obtain

P " p+ P ' ( 2 1 t 2 - o l * P ( o- Z - l ) : 0 ra 7l I

andthensubstituting the powerseriesexpansion of p andsolvingfor theai,swe

arriveat a recursionrelationfor the ai. coefficients:
(A-o*L+l)at
(2.71)
(r+l)(ft+21+2)
For largek, the ratio of expansioncoefficientsreachesthe limit ar,+r at : I k.
I I
which has the samebehavioras the power seriesexpansionof ep. Becausethe
70 M o d e l p r o b l e m st h a t f o r m i m p o r t a n ts t a r t i n gp o i n t s

po$'erseriesexpansion of P describes a functionthatbehaves like e/,fbr largep.

the resultingS(p) functionvr'ouldnot be nonnalizablebecausetlree-pr'l factor
would be overw'helmed by this e/' dependence. Hence.the seriesexpansion of p
must truncatein orderto achievea normalizable"! function.Noticethat ilp is
imaginarv.as it rvill be if E is in the continuum.theargumentthattheseriesrnust
truncateto avoidan exponentiallvdii,er-sing functionno longerapplies.Thus.
we seea key differencebetu'eenbound lwith p real) and continuum(u,ith p
imaginary)states.In the formercase.the boundarvconditionof non-dirersence
arises:in the larter.it does not becauseexp(p/2) does not divergeif p is
rmaginary.
T o t r u n c a t e a t a p o l y n o moi a
f ol r d e r n ' .w e m u s t h a vne, - o + L + I = 0 . T h i s
impliesthatthe quantityo introducedpreviouslyis restrictedto o- : n' t L * l.
which is certainlyan integer;let us call this integern. If u,elabelstatesin order
o f i n c r e a s i nng : 1 . 2 . 3 . . . . . w e s e et h a td o i n - s o i s c o n s i s t e nwt i t h s p e c i f y i n g
a r n a x i m u mo r d e r( n ' ) i n t h e p ( p ) p o l y n o m i anl ' : 0 . l . 2 . . . . a f t e ru , , h i c thh e
l-value canrun from Z : 0, in stepsof unity.up to I : n _ 1.
Substituting the integern for o, we find that the energylevelsarequantized
because o is quantized(equalto l):
- ' J

.f : _:-;__;
llL-c
{-./t }
:IT II.

and the scaleddistanceturnsout to be

Zr
(2.761
aorl

Here,the lengthas is the so-calledBohr radius(ao : li2 ,rnz);it appearsoncethe

1
aboveexpressionfor f is substitutedinto the equationfor p. Using the recursion
equationto solve for the polynomial'scoefficientsa1,for anychoiceof r and /
quantumnumbersgeneratesa so-calledLaguerrepolynornial;p,,_1_1(p).They
containpowers ofp fromzerothroughn - L - l. andtheyhaver - r - I sign
changesas the radial coordinaterangesfrom zero to infinity. It is thesesign
changesin the Laguerrepolynomialsthatcausethe radialpartsof the hydrogenic
wavefunctionsto haven - L - 1 nodes.For example,3d orbitalshaveno radial
nodes,but 4d orbitalshaveone;and as shownin Fig. 2.16.3p orbitalshaveone
while 3s orbitalshavetwo. once again,thehigherthe numberof nodes.the hieher
the energyin the radial direction.
Let me again remind you about the dangerof trying to understandquantum
wave functions or probabilitiesin terms of classicaldynamics.what kind
of
p o t e n t i azl ( r ) w o u l d g i v e r i s e t o , f o r e x a m p l e , t h e 3 s p ( r ) p l o t s h o w n i n F i g . 2 . l 6 ?
classical mechanicssuggeststhat p should be large where the particle moves
slowly and small where it movesquickly. So, the 3s p(r) plot suggeststhat the
radialspeedof the electronhasthreeregionswhereit is low (i.e.,wherethepeaks
in P are)andtwo regionswhereit is very large(i.e.,wherethe nodesare).
This.in
 H y d r o g e n i co r b i t a l s
t l

€
!

Plotsof the
r a d i a lp a r t so f t h e 3 s a n d
3p orbitals.

turn,suggests thattheradialpotentialz(r) experiencecl by the 3s erectronis high

in threeregions(nearpeaksin p) and low in two regions(and
at thenucreus). of
course.this conclusionaboutthe form of v(r) ls nonsense
and againillusrrates
how one must not be drai.vninto trying to think of the
classicalmotionof the
particle.especiallylor quantumstateswith small quantum
number.In fact,the
low quantumnumberstatesof suchone-electron atomsand ionshavetheirradial
P(r) plots focusedin re-qions of r-spacewherethe potential-Ze2
1r is most
attractive(i.e..lar_gest in magnitucle).
Finally.we notethatthe energyquantization doesnot arisefor stateslylng in
the continuumbecause the conditionthat the expansionof p(p\terminatedoes
n o ta r i s eT. h es o l u t i o ncsl f t h er a d i ael q u a t i o a
n p p r o p r i a treor h . s . s c a t r e r i nsgr a t e s
(which relateto the scattering motionof an electronrn the field of a nucleusof
chargeZ) area bit outsidethe scopeof this text. so we will
not treatthem further
here. For the interestedstudent,they are treatedon p.
90 ofthe text by Eyring,
Walter,and Kimball.
To review separationof variableshas been used to solve
the fulr (r. {), Q)
Schrridinger equationfor oneelectronmovingabouta nucleusof chargeZ.The
0 andQ solutionsarethe sphericalharmonicsyt.,,(0,@).The
bound-state radial
solutions

R , , . 7 Q ): S ( p ) = p L s - r ' / 2P , - r - r ( p ) (2.77\
dependon the ir and / quantumnumbersand are given in terms
of the Lasuerre
polynomials.
 tz M o d e l p r o b l e m st h a t f o r m i m p o r t a n ts t a r t i n gp o t n t s

2 . 5 . 4S u m m a r y
To sunrmarize. the quantumnumbersL and, nt arisethroughboundaryconditions
r e q u i r i n g t l t a t i l t (b0e) n o r m a l i z a b r e ( i . e . " n o t d i v ear n
g cel)t t r ( 4 t r : t / r ( 4 t t 2 r \ .
The radialequatio'.which is theo'ly pracethe potenriarenergy
e'ters.is found
to possess bothboundstates(i.e..statcswhoseenergieslie belou.theasynrptote
a t * h i c h t h e p o r e n r i ar a l r r i s h easn d t h e k i n e r i ce n e r g li s z e r t r c) ' d
conlirrruni
stateslying energeticalll, abovethis asynrptote. The resultinglryclroeenic u,ave
f u n c t i o n s( a n g u l aar n d r a d i a r ) a n de n e r g i eas r es u n r m a r i z e d p p .
on 1 3 3 - r 3 6i n
the text by Paulingand Wilsonforr up to anclincluding6
anclI up to -5.
There are both bound and continuumsorutionsto the radiar
Schrcidinger
equationfor the attractivecoulornb potentialbecausc.at cnelgies
bel,r.r,, the
asymptote. thepotentialconfinestheparticlebetweenr : 0 andan outer
turnin_9
pornl.whereasat energies abovetheasymptote. theparticleis no longerconfined
b v a n o u t e rt u r n i n gp o i n t ( s e eF i g . 2 . r 7 ) . T h e s o l u t i o n s
ofthis one-electron
proble'r fornr the qualitativebasisfor n.ruchof atomic
and nrolecularorbital
theory.For this reason.the readeris encouraged to gain a firr.rerunclerstandin_r
ofthe natureofthe radialandangularpartsofthese*,avefuncrions.
The orbitals
thatresultarelabeledb,vn. L. and,r quantur,nurnbersfor
the lroundstatesand
by z and,,r quantumnumbersandthe energy6 for the continuur.n
states.Much
as theparticle-in-a-bor orbitalsareusedto qualitatively clescribe z-electronsin
conjugatedpolyenes,theseso-calledhydrogen-like orbitalspro'ide qualitative
descriptions of orbitalsof atomswith nrorethana singleelectron.By introducing
the conceptof screeningas a way to represent the repulsiveinteractions among
the electronsof an atom,an effectivenucrearchargeZsl.ca.be
usedin placeof
Z inthe rfln.1.n,.<nd 8,, to generateapproximate atomicorbitalsto be filled by
electronsin a rnany-electronafom. For example,in the
crudestapproximation
ofa carbonatorn,the two rs electronsexperiencethe full
nuclearattractionso
Zen:6 for them,whereasthe 2s and 2p electror.rs are screened by the two ls

-r"a,,'l
0.0

'-"1
Radial
potentialfor hydrogenic
a r o m sa n d b o u n da n d
c o n t i n u u mo r b i t a l
energies.
 Electron
tunnelino 73

electrons,so 2"6 : 4 for them. Within this approximation,one then occupies

nvo 1sorbitalswithZ:6, two 2s orbitalswith Z:4 and two 2p orbitalswith
z - 4 in formingthe full six-electronwavefunctionof the lowestenersvstateof
carbon.

2.6 Electrontunneling
T u n n e l i n gi s a p h e n o m e n o n o f q u a n t u m m e c h a n i c s ,n o t c l a s s i c a lm e c h a n -
ics. lt is an extremely important subject that occurs in a wide variety of
chemicalspecies.

Solutions to the schrodinger equation display several properties that are very
different from what one experiences in Newtonian dynamics. one of the most
unusual and important is that the particles one describesusing quanfum mechanics
can move into regions of space where they would not be "allowed" to go if they
obeyedclassicalequations.Let us consicleran example to illustratethis so-called
tunneling phenomenon. Specifically, we think of an electron (a particle that we
likely would use quantum mechanics to describe) moving in a direction we will
call R under the influence ofa potential that is:

(i) Infinitefbr R < 0 (thiscoulclfor example.represent a regionof spacervithina

solid materialrvherethe electronexperiencesvery repulsiveinteractions,,vithother
electrons ):
(ii) constanta'd ne-eative for somerangeof R betrveenR : 0 and R,,,",(this could
represent the'attractiveinteraction
of the electronswith thoseatomsor molecules
in a finiteregionof a solid):
(iii) Constantand repulsiveby an amountAI/ + D" fbr anotherfinite region
frorn R*o,
to R,"a^+ d lthis could representthe repulsiveinteractionsbetrveenthe electrons
and a layerof moleculesof thickness6 lying on the surfaceof the solidat Rn,,");
(iv) constantandequalto D. fiom R,,,",+ 6 to infinity(this
couldrepresent the
electronbeing removedfrom the solid"but with a work functionenergycost of D",
and moving fieely in the vacuumabovethe surfaceand the ad-laver).Sucha
potentialis shorvnin Fis. 2. I 8.

One-
d i m e n s i o n apl o t e n t i a l
s h o w i n ga w e l l ,a
R.r* Rn,u"+ d b a r r i e ra, n d t h e
ElectronpositioncoordinateR + asymptoticregion.
74 M o d e l p r o b l e m st h a t f o r m i m p o r t a n t
s t a r t i n gp o i n t s

The pieceu'isenatureof this potentiarallows

the one-dinre'sionar Schrridi'ger
e q u a t r o n t o b e s o l v e d a n a l y t i cFaol lrye.n e r g i e s
l y i n g i n t h e r a n gD
e" < E < D"
6[', an especiall,r, interesting *
classof sorutions exrsts.Trieseso_cailed
srares resonance
occurat energies thataredete''ined by thecondition
thatthearrpritudeof
thewavefunctionwithinthebarrier(i.e.,for'0:
R < R,,,,,,)belarge.Letusno
turn our attentionto this specificenerg-v
,.gi,r',..i.i,hichalsoser\,,es to introcruce
the tunnelingplrenomenon.
The piecervise sorutionsto the Schrridinger equationappropriate to the reso_
nancecaseareeasirywrittendoil'nin
termsof sinand.o, or.*pon.ntiarfunctro's,
usingthe followingthreedefinitions:

k':\E;c-D\v,r. , -
r' : 11 + SI._ g117;.
2nr"(D"
( 2 . 7 )8
The c.mbinationsof sin(rtR)and
cos(AR)that sorvethe Schrcidin-eer
tn the innerregionand that vanish equario.
at R : 0 (because the functionntustvanlsh
u i t h i n t h er e g i o nw h c r e/ ' i s i . f i n i r e
a n db c c a u s lcr n l u s rb e e o n r i n u o u s .
r''anish at R : 0) is ir musr

W:..1 sin(/,R) (for0< R S Rn,"*).

(2.79t
BetweenR'r^ and Rnr"^* 6. there
are two sorutionsthat obey trreSchrodinger
equation'so the mostgeneralsorution
is a combinationof thesetwo:
W : B n e x p ( r ' R*)B exp(_r,R) ( f o r R n ,<
o .R < R n , u , + d ) . (2.g0)
Finally.in theregionbeyondR,ro"
* d,wecanusea combination of eithers in(k,R)
and cos(fr'R)or exp(ii'R) and exp(-i
t'R) to expressrhe sorurion.Unlike
regionnearR : 0. whereit u as most the
convenient to usetrresin andcosfunctrons
becauseoneof therncould be "thrown
away"sinceit courdnot meet
c o n d i t i o n o f v a n i s h i n g a t R : 0 , i n t h i s l a r y e - R r e g r o n , e r t h e r s ethe
t i sboundary
acceptable.w
chooseto usetheexp(i/,'R) andexp(-i
t"i,, setbe.auseeachofthesefunctrons
an eigenfunction rs
of the momentumoperator-ih,-
/a R.This alrorvsus to discuss
amplitudesfor electronsmo'ing
with positive momentum and with
momentum.So. in this region,the negative
most generalsolutionis
V : Cexp(ll'R)+ Dexp(_if,R) (forR,o+
, 6 < R < cc).
(2.81)
Therearefour amplitudes(A, B+
^ ,8-, andC) thatcanbe expressed in terms
of the specifiedamplitude D of the
incoming iu^ 1..g., pretendthat
the flux of electronsthat our we know
experimentalulpuru,r, ..shoots,,
Four equationsthat can be used at the surface).
io achievethis goarresult when
arematchedat Rmu^and at R-u* v and dv/d^
* 6 (oneof the essentiar propertiesof solutions
to the Schrcidinger equationis that they and their flrst
derivativeare continuous;
thesepropertiesrelareto V
beinga profabilitf un._, rU
1U Obeinga momentum
 Electron
tunneling 75

operator).Thesefour equationsare
.4 sin(tR.".) : B-. exp(r'Rn'u.)* B exp(-r'R,',,*), ( 2 . 8)2

It cos(l'R"','): t<'B* exp(r'R..,) - rc'B- exp(-r'R.o*) ( 2 . 8 )3

B - e x p ( r ' ( R , n "+- t ) ) + B - e x p ( - r ' ( R , u * * 6 ) )

: C e x p ( r t ' ( R n , ,*r 3 ) + D e x p ( - i t ' ( R n , o** 3 ) , (2.84)

r ' B - e x p ( r ' ( R . u * 6 ) ) - r ' B - e x p ( - r ' ( R , n " ,+ 6 ) )

= ik'C exp(ik'(Rn,"* * 3)l - ik D exp(-i /r'(R",",* 6)). (2.85)

It is especiallyinstructiveto considerthe valueof A I D thatresultsfrom solving

this set of four equationsin four unknownsbecausethe modulus of this ratio
providesinfbrmation about the relative amount of amplitudethat exists inside
the barrier in the attractiveregion of the potentialcomparedto that existing in
the asymptotic region as incoming flux.
The result of solving for .4I D is

* 3 )] {exp(r'3 )(ik' - r')[r' sin([ Rn,,,,

A I D : ;lr,,'exp[-1,{'(R,,,,,, )
+ I c o s ( , ( R , , , " .i)k)'l + e x p ( - l r ' l i ) ( i t ' + r ' ) [ r ' s i r r ( R
l ,,,,)
- I cos(/r/?,,,,,, ' ( 2 . 8)6
))l i k' | .

Further.it is instructil'eto considerthis resultunderconditionsof a high (lar-ee

D. * d L' - E\ and thick (larged) barrier.In sucha case.the "tunnelingfau^tor"
exp(-r'd) lvill be i,'erysmallcomparedto its counterpartexp(k'6).and so
ik'r'
' 1I D : 1 - - L
-
e r p [ - i ( ' 1 R , , '+. , ') ) ]c x p -( r ' 6 ) [ r ' s i n ( k R , ' 1
".
tA r'
'
+ ,( cos{{'R,,,.,,
)l t 1 . 8r7

The exp(-r'd) factorin A I D causesthe magnitudeof the rvavefunctioninside

the barrierto be small in nrostcircumstances; we say that incidentflur must
tunnelthroughthe barrierto reachthe innerregionand thatexp(-r'd) -eives the
p r o b a b i l i toyf 't h i st u n n e l i n u .
Keepin rnindthat. in the energvrangewe are considering(E < D" * d), a
classicalparticlecould not evenenterthe region R,n".{ R . R,..* f d; this is
why we call thistheclassically forbiddenor tunnelingregion.A classicalparticle
startingin the large-Rresioncan not enter.let alonepenetrate. this region,so
sucha particlecouldneverendup in tlie 0 < R < R,.,,,innerregion.Likewise.a
classical particlethatbeginsin the innerregioncanneverpenetrate thetunneling
regionandescapeinto the large-Rregion.Wereit not for the factthat electrons
obeya Schrodinger equationratherthan Newtoniandynamics.tunnelingwould
not occurand for example.scanningtunnelingmicroscop;- (STM), ll'hich has
provento be a wonderfuland pow'erfultool for imaging moleculeson and near
 76 M o d e lp r o b l e m st h a t f o r m i m p o r t a n ts t a r t i n g
points

surfaces. wourdnot exist.Likewise,manyof the

devicesthatappearin our mod_
ern electronictoolsand which dependon currentsinducecr
-qalxes, by tunneiing
throughvarious-iunctions, wourdnot be availabre.But. rf .;;;;;.,rn*
occurand it can haveremarkableeffects. oo.,
trt. ur examinean especia,yimportant(in
. chemrstry,)phe'omenonthattakes
placebecauseof tunneringand that
occurswhen the energyE assumesvery
values.The magnirudeof the A/D
:p:cial factorin the abovesorutionsol.the
Schrcidinger equationcan becontela.g. ifth. energy
C is suchthat
r srn(A'Rn,r,)
* A cos(kR,,,.*
) C . 8 8)
r s s r n a l lI.n f a c t .i f

t a n ( { ' R , r , r: " ) - k l r '

(2.89)
thedenominatorfactorin AID will vanish andA/D will
becomeinfinite.lt can
be shownthattheaboveconditionis similar
to the energyquantization
condition
tan(lR,n,*): -k/r
(2.e0)
that ariseswhen boundstatesof a finite
potentiarwe, are examined.There
hou'ever.a difference.In thebound-state is,
situation,rwo energy-reratecr
paranreters
occur

k: \EpE/tt ( 2 . e) r
and

,:rEu1o"-syy; (2.92)
In the casewe are now considering,fr
is the same.but

( : . 9 3)
ratherthanr occurs,so the two tan(frR.u")
equationsare not identical,but they
are quite similar.
Anotherobservationthat is usefurto
makeaboutthe sifuationsin whichA
becomesvery largecanbe madeby considering /D
thecaseofa very high barrier(so
that r' is much larger thank).In this
case,the lenominator that appearsin
A /D.
_R' m a x
r' sin(tR."^)* I cos(lR.u*): r, sin(rtR,u,),
Electron coordinate R
_ (2.e4)
canbecomesmallif
One-
m e n s r o n apl o t e n t i a l
s i m i l a rt o t h e t u n n e l i n o sin(tR-u,,;:9.
(2.es)
pote.ntialbut without
t-he This conditionis nothing but the
oarrrerand asymptotic energyquantizationconditionthat occurs
the particle-in-a-box for
region. potentialshown. nig. z.re. This
potentiaris identicalto
the potentiarthat we were examining
for 0 5 R < Rma*,but extendsto infinitv
 E l e c t r o nt u n n e l i n q

beyondR,n,,, ; thebarrierandthe dissociationasymptotedisplayedby our potential

areabsent.
Let's consider'"vhatthis tunneling problem has taught us. First, it showed
us rhatquantumparticlespenetrateinto classicallyforbiddenregions.It showed
that.at certainso-calledresonance tunnelingis muchmorelikelythanat
energies,
"off resonance".In our modelproblem,
ener-sries thanare this meansthat electrons
impingingon thesurfacewith resonance energieswiil havea very high probability
of tunnelingto producean electronthat is trappedin the 0 < R < Rn,o"resion.
By the wav. we could have solved the four equationsfor the amplitude C
of the outgoing wave in the R ) R.o* region in terms of the I amplitude.We
might want to take this approachif wantedto model an experimentin which the
electronbeganin the0 < R . R,,,".regionandwe wantedto computetherelative
amplitudefbr the electronto escape.However,if we were to solve for C lA and
then examineunder what conditionsthe amplitudeof this ratio w'ouidbecome
small (so the electroncan not escape),we would find the sametan(ftRma,):
-k f rc' resonance conditionaswe fbund from the otherpoint of view This means
that the resonance energiestell us fbr what collision energiesthe electronrvill
tunnel inward and producea trappedelectronand, at these same energies,an
electronthat is trappedwill not escapequickly.
Wheneverone has a barrier on a potentialenergysurface.at energiesabove
the dissociation asymptotcD. but belowthe top of the barrier(D. + ,l y' here),
onecan expectresonance statesto occurat "special"scatteringenergiesE. As
we illustratedrvith the model problem,theseso-calledresonance energiescan
oftenbe approxirnated by the bound-state energies ofa potentialthat is identical
to the potentialof interestin the innerregion(0 < R < R',"*)but that exlends
to infinitybeyondthe top of the barrier(i.e.,beyondthe barrier.it doesnot fall
b a c kt o r a l t r e sb e l o u E ) .
The chemicalsignificance of resonances is great.Highly rotationallyexcited
moleculesrray ha'"'emore thanenor.rgh total energyto dissociate(D"). but this
energymaybe"stored"in therotationalmotion.andthevibrationalenergymaybe
lessthanD.. In termsof the abovemodel.high angularmomentummayproduce
a significantcentrifugalbarrierin the effectivepotentialthat characterizes the
molecule'svibration,but the system'svibrationalenergymay lie significantly
below D.. In sucha case.and when viewedin termsof motion on an angular-
momentum-modified efi'ectivepotentialsuchas I showin Fig. 2.20.the lifetime
of themoleculewith respectto dissociation is determined by therateof tunneling
throughthe barrier.
In that case,one speaksof "rotationalpredissociation" of the molecule.The
Iifetimer canbeestimated by computingthefiequencyu at whichflux thatexists
insideRrr* strikesthe barrierat R,,,,,,,
t1K
y = _ (s ,) (2.e6)
2trR,",,,
 78 M o d e l p r o b l e m st h a t f o r m i m p o r t a n ts t a r t i n g
potnts

Melrstcblc rotJtronal le\ el

rG) + h?[Ju +

Radial
potentialfor
l'(R) for non-rolatingmolecule
n o n - r o t a t i n g( J : 0 )
motecule and for
rotating molecule.

and thenmultiplyingby the probability p

rhatffux tunnelsthroughthe barrier
fronr R.", to R,rr^* d:

P = exp(-2r'3) (2.e7)
The resultis that

t l N
f = --- c' "\ F
n /t - ) , , '' ^t r u l
) " P | InlSr
-ts
(2.e8 )
u'ith the energy f entering into fr
and r' being determinedby the resonance
c o n d i t i o n(:r ' s i ' ( t r R - o ' ) * r c o s ( r R - , " ) ):
m i n i m u m .e y l o o t i n g b a c ka t t h e
defintionof r', we note that the probability
of tunneringfalrs off exponentia'y
with a factor dependingon the rvidth d
of the barrier throughwhich the particre
must tunnelmultipliedby r,, which depends
on the heightof the barrierD" 3
abovethe energy .E available.This exponential *
dependenceon thicknessand
height of the barriersis somethingyou
should keep in mind becauseit appears
i n a l l t u n n e l i n gr a t ee x p r e s s i o n s .
Another importantcasein u,hich tunnering
occursis in erectronicailymera_
stablestatesof anions.In so-calredshap..einunce
srares,,n.-r"t""i',,.^o",,
electronexperiences:

(i) anattractivepotentiar
dueto its interaction
with theunderlying neutralmorecure,s
dipole.quadrupole. andinducedelectrostaticmoments, aswellas
(ii) a centrifugarpotentialof.theformL(L + r)h2/gn r,"R2 *hos.
.agnitude
depends on theangurar characterofthe orbitaltheextraerectron occupies.
when co'rbined.the aboveattractive
and centrifugarpotentialsproducean
tive radial potentiarof the-form shown efrec_
in Fig.2..2rfbr the 1y'f casein which
addedelectronoccupiesthe n* orbital the
whii has z - 2 characterwhen vieu,ed
 E l e c t r o nt u n n e l i n g

Effective
r a d i a lp o t e n t i aflo r t h e
excesselectronin N!
o c c u p y i n gt h e z * o r b i t a l
w h i c hh a sa d o m i n a n t
L = 2 component,
L ( L + 1 l l ] / 8 n 2m # .

from the centerof the N-N bond.Again. tunnelingthroughthe barrierin this

potentialdetennine.s the lif-etimes
of suchshaperesonance states.
Althoughthe examplestreatedaboveinvolvedpiecewiseconstantpotentials
(so the Schrridinger equationand the boundarymatchingconditionscould be
solved exactly),many of the characteristicsobservedcarry overto morechenti-
callyrealisticsitr.rations.
In fact.onecanoftenmodelchemicalreactionprocesses
in termsof motionalonga "reactioncoordinate"(.r)froma regioncharacreristic
of reactantmaterialsr''utere the potentialsurfaceis positivelycurvedin all direc-
tion and all forces(i.e..gradientsofthe potentialalongall internalcoordrnates)
vanish.to a transitionstateat r.vhichthe potentialsurface's cllrvaturealon-q.r'
is R e g i o no f
a c t iV a t c d
negatir,e while all othercurvaturcsarepositir'eand all forcesvanish:onw'ardto
corrplex
productmaterialsrvhereagainail curvatures are positiveand all forcesvanish. Reactants
A prototypicultracc of the encrgyvariationalong such a reactioncoordinate
is sho'uvnin Fig.2.22.Near the transitionsratcat the top of the barrieron this Energy
surface.tunnelingthroughthe barrierplaysan importantrole.if the massesof ile along a reaction
p a t hs h o w i n gt h e b a r r i e r
the particlesmoving in this regionare sufficientlylight. Specifically. if H or D t h r o u g hw h i c ht u n n e l i n g
atomsare involvedin the bondbreakingandformin_e in this regionof the cncrgy m a y o c c u r .
surface. tunnelingmustusuallybe considered in treatingthe dynamics.
Withinthe above"rcactionpath"pointof view..motiontransvcrse to the reac-
tion coordinate.i is clftenniodeledin termsof local harmonicmotionalthough
moresophisticated treatmcntsof the dynamicsare possible.This pictureleads
one to considermotion along a singlc degreeof freedom(s), with respectto
tvhichnruchof the abovetreatmentcan be carriedover.coupledto transverse
motionalongall otherinternaldegreesof freedomtakingplaceunderan enrirely
positivelycurvedpotential(which thereforeproducesrestoringfclrcesto mole-
mentaw'ayfrom the "streambed"tracedout by the reactionpaths ). This point of
view constitutes oneof the mostwidelyusedanclsuccessful modelsof molecular
reactlondynamicsand is treatedin moredetailin Chapter8 of this text.
80 M o d e lp r o b l e m st h a t f o r m i m p o r t a n ts t a r t i n gp o i n t s

2 . 7 A n g u l a rm o m e n t u m
2 . 7 . 1O r b i t a a
l n g u l a rm o m e n t u m
A particlenro'ing rvith momentump at a positrclnr relative
to sonrecoordr-
nateorigin has so-calledorbital angularmomentumequal
to L : r x p. The
threeconrponents of this angularmomentumvectorin a cartesiancoordinate
systemlocatedat the origin rnentioned
abovearegivenin termsof the cartesian
coordinatesof r and p as follo,,vs:

L-=tp,.-tJt,. ( 1 . 9)9
L, : . t ' l t .- : 1 t , . ( 2 .I 0 0 )
L,.::p, * xlt,.
( 2 l. 0 l 1
using the fundanrentar
commutationrerationsarlong the cartesiancoordi-
natesandthe Cartesian
momenta:

[ q t . p i l : 4 t p r - p i 4 t = i h 6 /1 i . t = , r . r , . : ) .
tl.l0lr
rt canbe shownthatthe abo'e angularmomentum
operators
obeythe folro$,ing
s e lo f c o m m r r t a t i or n
elations:

[ 2 , .Z , ] - i t t L . . ( 2 r. 0 3)
lL,.L.l: itlL,. (2.r0.1)
lL,.L,l: ihL,. (2.105)
Although the cornponentsof L do rot comnlutewith
one another.they can be
shoivnto commutewith the operatorZ2 definedby

L) : L: + Li + L:. (2.106)
This new operatoris referredto as the squareof the
total angularmomentum
operator.
The conmutation propertiesof the components
of L allow us to conclude
that completesetsof functionscan be founcithat are
eigenfunctionsof Z2 and of
one, but not more than one, componentof L. It is convention
to selectthis one
componentas 2., and to label the resultingsimultaneous
eigenstates of L2 and
L, as ll. n) accordingto the correspondingeigenvalues:

L2lt.nt):hlt(t+l)ll.n), /=0. 1.2.3.....

(2.107)
L , l l . n ): l a n l l , m ) , m : * t , * ( / - t ) .+ ( t _ 2 ) . . . . t ( t _ ( / _
I)).0
( 2 .r 0 8 )
Theseeigenfunctionsof L2 andof z, will not, in
general,be eigenfunctionsof
eitherz' or of 2... This meansthat any measurement
of z, or 2.,.wilr necessariry
changethe uave function if it beginsas an eigenfunction
of 2,.
The aboveexpressionsfor Z._,L,, andL, canbe
mappedinto quantumrne_
chanicaloperatorsby substitutingx,
v-, and z as the correspondingcoordinate
 A n g u l a rm o m e n t u m 81

operatorsand-i/10 l0x . - ihd ldt,, and -ihd l0z for p.,. p,,, andp_-.respectively.
The resultingoperatorscan then be transformedinto sphericalcoordinatesthe
resultsof which are

L, = -uto/tte. ( 2 .l 0 e )
L . : i t t l s i n Qi ) l A e+ c o r 0c o s gd l i ) Q | . ( 2 .l l 0 )
L , : -i fi{cos d AlAe - cot0 singdl0Ql. (2.1 I t)
L 2 : - t 1 : { ( l /s i n 6 )d / a d ( s i n d l a 0 +
) ( l / s i n :0 ) d 2l , ) O 2 l ( 2 . rl 2 )

2 . 7 . 2P r o p e r t i e so f g e n e r a la n g u l a rm o m e n t a
There are nlany types of angular momentathat one encountersin chemrsrry.
orbital angularmomenta.snchasintroduced above,arisein electronicmotionin
atorrs.ln atom-atomand electron-atom collisions,and in rotationalmotion in
molecules. Intrinsicspinangularmomentumis presentin electrons. Hl , H2,cl3.
andmanyothernuclei.In this section,we will dealwith the behaviorof anvand
all an_qular momentaandtheircorresponding eigenfunctions.
At times.an atom or moleculecontainsmore than one type of angularmo-
mentum.The Hamiltonian's interaction potentialspresentin a particularspecles
may or may not causetheseindividualangularmomentato be coupledto an
appreciable extent(i.e..the Harniltonianmay or may not containtermsthat re-
t'ersimultancously to two or moreof theseangularmomenta).For example.the
NH- ion. lvhichhasa rf[ grouncielectronicstate(its electronicconfigurarion is
1 s i , 2 o l 3 o r 2 p ; , : p l ,) h a se l e c t r o n iscp i n .e l e c t r o n iocr b i t a l a. n dm o l e c u l arro r a -
tionalangularmomenta.The full HamiltonianH containstermsthat couplethe
electronicspin and orbitalangularmomenta.therebycausingthem individuallv
to not commutervith H.
In suchcases.the eigenstates of the systemcan be labeledrigorouslyonly
by angularmomenturnquanrumnumbersj andn belongingto the totalangular
momentumJ. The totalangularmomentumof a collectionof individualangular
momentais defined.component-by-component. as follows:

"l '. -- \ - t - Itli). ( 2 . I13 )

where,tlabels.y.-r,,and:, andl labelsthe constituents

whoseangularmomenra
coupleto produceJ.
For the remainderof this section.we will study eigenfunction-eigenvalue
relationshipsthat are characteristic
of all angularmomentaanclwhich are con-
sequences of the commutationrelationsarxongthe angularmomentumvector's
threecomponents. we will alsostudyhow one combineseigenfunctions of two
or moreangularmomenta{J(i )} to produceeigenfunctions of the total J.
82 Modelproblems
thatform important
startingpoints

Consequencesof the commutation relations

Any setofthreeoperators
thatobey

[J,. J,] - ittJ,. (2.1t1)

14. .I.l : ittJ,. (2.1 l-5)
[J-.J,l : irlJ,. ( 2 r. l 6 )
will be takento definean angularmomentunt
J. whosesquareL) : Ji + Ji + .l;
commuteswith all three of its components.It is useful to also introducetwo
combir.rationsof the threefundamentaloperators:

J*:.1, tiJ., Q.lt7)

and to refer to them as raising and lowering operatorsfor reasonsthat r,r,illbe
made clear below. These ne\\' operatorscan be shown to obel, the follorving
commutationrelations:

l J 2 .J - l = o . ( 2 . 1l 8 )
1.1.. Jxl - tTtJ+. (2.1l9)

Usingonly theabovecommutation properties,it is possibleto proveirnportant

propertiesof theei-cenfunctionsandeigenvalues of J2 andJ.. Let us assumethat
lvehavefounda setofsimultaneous eigenfunctions ofJ2 andJ.: thefactthatthese
two operatorscol.nnrute
tells us thatthis is possible.Let us labelthe eigenvalues
belongingto thesefunctions;

J . l 1 . m 7: n ' . f t i . m ) l j . n t ) . (2.120)
J . l j . n t )- h n l j . n ) . ( 2 . 1I2)
in termsof the quantities
rl andJ'(j . m). Althoughwe cer.tainly "hint"
thatthese
quantitiesmust be relatedto certainj andnt quantumnumbers.we havenot yet
proventhis, althoughwe will soondo so.For now.rvevieu' ./'(.1. m) andrir simpll,
as symbolsthat representthe respectiveeigenvalues. Becauseboth J2 and J, are
Hermitian,eigenfunctions belongingto different.f ( j. m) or r? quantumnumbers
must be orthogonal:

( i . n t I j . m ' ) = 3 , , , . , , ., , 5 (2.122)
we now prove severalidentitiesthat are neededto discoverthe inforrnatron
about the eigenvaluesand eigenfunctionsof generalangularmoment3that we
are after.Later in this section,the essentialresultsare summarized.

(i) There is a maximum and a minimum eigenvaluefor ,I,

Becauseall of the componentsof J areHermitian,andbecausethe scalarproduct
of any function with itself is positivesemi-definite,the following identity holds:

( j . n t l J i + L l . t . i . m l : \ J , ( . j , m l J , u . m+) ( J , \ j , n r l J , ) j . n )> 0 . (2.t23)

 A n g u l a rm o m e n t u m 83

-
Ho*,ever.Ll + li is equalto J2 -/-2,so this inequalityimplies that
( j . m l J ) - l i l i . m ) = / r 2 1 1 ' 1nit.1- m 2 l > 0 . (2.1211)

nhich. in turn, implies that nt2 must be lessthan or equalto -f(j.lz). Hence,
for any value of the total angularmomentum eigenvalue/, the ;-projection
eigenvalue(rr) must havea maximum and a minimum value and both of these
mustbe lessthan or equalto the total angularmomentumsquaredeigenr,aiue ,/'.

(ii) The raising and lowering operators changethe "I. eigenvaluebut not
the ,I2 eigenvaluewhen acting on 17.rl)
Applyingthe commutationrelationsobeyedby Ja to 17.z ) yields anotheruseful
result:

J , J " l j . n ) - f i J , l j . m ) : I l 1 1 o11. . 7 , (2.t2s)

J 2 J L l j . m -) J * J ) l i . n t ) - 0 . (2.126)

Now,usingthe fact that 17.n) is an eigenstate

of J2 and of -/,, theseidentities
give

J . J - \ j . n ) : 1 p 7+ l i ) J = l j . n l = l d m + l ) l j , n ) . (2.t27)
Jt Jt .l .n) : r11
.11i.nJ J-) j. m'1. ( 2 .r 2 8 )

Theseequationsprovethat the functionsJalj. m) must either themselves be

e i g e n f u n c t i o n s o f - / r a n c i - / . , w i t l i e i g e n v a1l u1e. 1.sm
n2) a n d h ( m- t l ) o r J + l j . m )
mustequalzero.In the formercase.we seethat I actingon Ij, n) generates
a new eisenstatewith the srme Jr eigenvalue as lj. m) but with one unit of fi
higheror lolver in -/, eigenvalue.It is for this reasonthat we call .,/* raisingand
lorveringoperators. Noticethat,althoughJalj. m) is indeedan eigenfunction of
. r . w i t h e i g e n r , a l u( em t l ) l i . J + lj . i r r ) i s n o t i d e n t i c atlo l j . m t 1 ) ; i t i s o n l y
proportionalto l/. r? + l):

J - l i . m ) : C 7 , , , l jn. t J : 1 ) . ( 1 .I 2 9 )

Explicitexpressions fbr theseCf,,,coefficients will beobtainedbelow.Noticealso

that becausethe -/- 17.rr ) , and hence17.m X. 1), havethe same-I2 eigenvalueas
17.n) (in fact.sequential applicationof -Q canbe usedto showthat all lj, m'),
for all nr'. havethis sameJ2 eigenvalue). the -/2 eigenvaluef(j.n) must be
tndependent ofrr. For this reason,./ canbe labeledby one quantumnumber7.

(iii) The ./2 eigenvaluesare related to the maximum and minimum ,I-
eigenvalues which are relatedto one another
Earlier,we shor.vedthat thereexistsa maximumand a minimum value for n. for
any glven total angularmomentum.It is when one reachestheselimiting cases
84 M o d e l p r o b l e m st h a t f o r m i m p o r t a n ts t a r t i n gp o i n t s

that J-l.i . m) : 0 applies.In particular,

J * l ..jn t ^ ^ ) : 0 . (2.r30)
J-l j. nt^i,) = Q. ( 2r. 3 1
)
Applying the following identities:

J J. - J2 - J: -hJ,. (2.132)
J+J : J2 - J.2+h.1,, ( 2 .r 3 3 )

respectively.to 1.7,z-u*) and 17.nln ln) gives

r / l . f j . r r - a " ) - r r l o *- n r * o , ]: 0 . (2.134)
h1l.f U. /??min) - zrfl,n+ rr,ln = o. ( 2 . 31 5)
)
which imrnediately givesthe J2 eigenvalues
f(7. n.r^) and.f (.j. ntni) in terms
of r?n,u*or mmin'.

f (.j. m^^,): nl.u*(riro^ * I ). ( 2 .I 3 6 )

. l (j , m ^ , " ) : r l r n , ; n ( n l n-, ; nl ) (2.1
37)

So, u'e now know the -I2 eigenvalues for lj,m,no*)and 1.7.
rr?min).However,we
earliershowedthat 17.rr) and lj.m - 1) havethe same_/l eigenvalue (u,henwe
treatedthe effectof I on l.l,nr)) and that the J2 eigenvalueis independent
of
m. If we thereforedefinethe quantumnumber7 to be tnmar,we seethat the J:
eigenvalues aregivenb1

J 2 l i . m ) : t 1 2j ( j * t ) 1 1 , m 1 . (2.r38)

We also see that

-f(j,,n): j ( j * 1 ) : m ^ u * ( m ^ u+* 1 ) : n . ; n ( r l n , ; -n l ) , (2.13e)

from which it follows that

lzlmrn: -l/lmax.
(2.r40)

(iv) The j quantum number can be integer or half-integer

The fact that the z-values run from j to - j in unit steps(becauseof the properry
ofthe -! operators)meansthat thereclearlycan be only integeror half-integer
valuesfor 7. In the former case,the rfi quantumnumber runs over -j, -j +
l , - j + 2 , . . . , - j + ( j - 1 ) ,0 , 1 ,r , . . . , j t i n t h el a t r e r , , r ?
u n so v e r- 7 . - 7 *
l , - j + 2 . . . . , - j + ( i - l l 2 ) , 1 1 2 , 3 1 2 , . . . , 7 .O n l yi n t e g ear n dh a l f - r n r e g e r
valuescan range from j to -j in stepsof unity. Specieswith integerspin are
known as bosonsand thosewith half-integerspin are calledfermions.
 A n g u l a rm o m e n t u m AR

(v) Moreon J+lj.m)

Usingthe aboveresultsfor the effect of -[ actingon Ij. m) and the fact that -/*
andJ- areadjointsof one anothel allows us to write:
( j . m l J - J * l j . n ) : | i . * l ( J ' - J : - h J , )U , m l
= t l 2 U U* t ) - m ( m* t ) ]
: l J + U ,m l J l l j , r y ) : ( C i ) 2 , ( 2 . 111)

whereCl.u,is the proportionalityconstantbetweenJ+U , m) and the normalized

functionIj , m + 1). Likewise,the effect of -I- can be expressedas

( j . m J a l _ l j , m )= q i , ^ l ( J . - J : + h J , ) t j . m )
: h 2 l i( i * t ) - m ( n - 1 ) |
: ( J _ ( j .m l J l j . m ) : ( C 1 . ) 1 , (2.t42)

whereC,..,,is the proportionalityconstantbefweenJ_l j, m) and the normalized

l i . m- t ) .
Thus,we can solve for C1.,,after which the effectof -/a on Ij . m) ts given by

J - l j . n t ) : t t l j ( j* l ) - m ( m + l ) ) ' / ? l m
/.t1l. (2.113)

2 . 7 . 3S u m m a r y
The aboveresultsapply to any angularmomentumoperators.The essentialfind-
ingscanbe summarized as follows:

(i) J) andJ. ha'e complete setsof simultaneous

eigenfunctions.
we labelthese
eigenfunctionsj. n): theyareorthonormalin boththeirm- and i -typeindices:
U , m j ' . n ' ) : 3 , , , .5, ,.,, .
(ii) These17,rr) eigenfunctions obey

Jtli..) :htjtj + I)li.m). / : integerorhalf-integer (2.1,14)

J.lj.ntl:tttnlj,m), , n = - j , i n s t e p s o fI t o * 7 . (2.145)
(iii) The raisingand lowerin-eoperators.-/+act on lj. m) to yield functionsrhat are
eigentunctionsof ./: with the sameeigenvaluesas Ij , m) and eigenfunctionsof -/,
with eigenvaluesof (rr * l)ft:

J . tj . n 1 : l t l j t j + l t - m t m * l ) l r r . i . m = l l . 1) lJ6 |

(tv) When Ja actson the "extremal"states17,j) or

lj, -7), respectively,the resultis
zero.
The resultsgiven aboveare,as stated,general.Any and all angularmomenta
havequantummechanicaloperatorsthat obey theseequations.it is conventionto
designatespecifickinds ofangular momentaby specificletters;however,it should
be kept in mind that no matterwhat lettersare used.thereare operators
correspondingto J2, J,, and I that obey relationsas specifiedabove,and there
do Model problems that form important starting points

are eigenfunctionsand eigenvalues that haveall ofthe propertiesobrainedabove.

For electronicor collisionalorbitalangularmomenta,it is commonto usel.l and
l.: for electronspin,S: and S, areused;for nuclearspin 1: and I aremost
corltlolt: and for molecularrotationalangularmontentum.N: and ,\; are ntost
common (althoughsontetimesJ2 and.t, nravbe used1.Whencvertwo or more
angularmomentaarecombinedor coupledto producea "total" angular
monrentum. the latteris desicnated
bi Jr and J .

2 . 7 . 4C o u p l i n go f a n g u l a rm o m e n t a
If the Hamiltonianunderstudy containstermsthat coupletwo or ntore angu-
lar momentaJ(i ), then only the componentsof the total angularmomentum
J : L J(i ) and J2 will commurewith f1. It is thereforeessenrialto label the
quantumstatesof the systen,by the eigenvaluesof J- and J: and to coltstruct
variationaltrial or model wavefunctionsthat are eigenfunctionsof thesetotal
angularmomentumoperators.The problemof angularmolrentum coupling has
to do with how to combineeigenfunctions of the uncoupledangularmomen-
tum operators,which are given as simple productsof the eigenfunctionsof the
individualangularmotnentafl;l.ii, m1).to form eigenfunctionsof J2 and J..

Eigenfunctions of J,
Becausethe individualelementsof J are formedadditively,but J: is not. ir is
straightforwardto form eigenstates
of

1.:lJJit; (2.t4'7)
i

simpleproductsof the form fft Ij i , nti) areeigenfunctionsof I

r . n t i n t ): l r , ( t ) f ]t . i i , m=i))* , * f l tr,.,,,). (2.148)

and have J, eigenvaluesequal to the sum ofthe individual ntpll eigenvalues.

Hence,to form an eigenfunctionwith specifiedJ andM eigenvalues, one must
conrbineonly thoseproductstatesll, lji,nti) whosen;fi sum is equalto the
specifiedM value.

Eigenfunctions of J2: the Ctebsch-Gordon series

The taskis thenreducedto forming eigenfunctions
lJ. M),given particularvalues
for the {jr } quantumnumbers.When couplingpairs of angularmomentall.i , m)
and lj', rr')), the total angularmomentumstatescan be written, accordingto
what we determinedabove.as

lJ,M) = lcl I,,,,,li. m)lj',nt'), (2.149)

#r
i,:
:.t A n g u l a rm o m e n t u m 6t

wherethe coefficientsCl.j!,i ., are calledvector coupling coefficients(because

angularmomenfumcoupling is viewed much like adding two vectorsj and j'
to produceanothervectorJ), and wherethe sum overm andm' is restrictedto
rhos!-terms for which ffi * m' : L'l. lt is more common to expressthe vector
couplingor so-called Clebsch-Gordon (CG) coefficients
as (j . m; j' . ni I J. II)
view them as elements of a "matrix"
and to whose columns are labeledby the
coupled-state-/. M quantum numbers and whose rows are labeled by thequantum
characterizing the uncoupled "product j'
numbers basis"7, m; ,m .It turnsout
thatthis matrix can be shownto be unitary so that the CG coefficientsobey

j ' , n l I J . t t 4 l . ( j . mj ';. m ' I J ' . M ' ) : d L . L , | u . , v , (2.I s0)

l{l.,r,

and

\-l ; -. ;' n' I r ,r.t\ j',m' ( 2 . 1 5) 1

/J\r\,.,J t\j.m: I J, ill)- :3u.,,6,,.,,,

This unitarity of the cG coefficientmatrix allows the inverseof the relation

giving coupledfunctionsin terms of the productfunctions:

(2.152)

to be r.vrittenas

l i . m l l j ' . n i ) : L U . n t :j ' . m ' I J M ) " ) J . M )

:
f t t . t 4 t ,i . m ; j ' m ' ) l J . I t ) . ( 2r.s 3 )

Thisresultexpresses
the productfunctionsin termsof the coupledangularmo-
mentumfunctions.

Generation of the CG coefficients

The CG coefficientscan be generatedin a systematicmanner;howeverthey can
alsobe lookedup in bookswherethey havebeentabulated(e.g.,seeTable2.rl
of R. N. Zare..lngulur illomentunr.John Wiley, New York (1988)).Here,we
will demonstrate the techniqueby rvhichthe CG coe{fcientscan be obtained.
but we rvill do so lor ratherlimited casesand refer the readerto more extensir,e
tabulations.
The strategywe take is to generatethe lJ, J) srate(i.e., the statewith ma.xi-
mum M-value) and to then use -/_ to generatelJ. J - l), after which the state
l./ - l. -r - l) (i.e.,the statewith one lower-/-value)is constructed by findinga
combination of theproductstatesin termsof which lJ, J - l) is expressed (be-
c a u s e b o tl hJ , J - 1 ) a n dl - r - l , J - l ) h a v e t h e s a mM e - v a l u eM : J -ll
w h i c hi s o r t h o g o n at ol l J . J - l ) ( b e c a u sleJ - 1 . . / - l ) a n d l J , - / - l ) a r e
88 M o d e l p r o b l e m st h a t f o r m i m p o r t a n ts t a r t i n gp o i n t s

eigenfunctions of the HermitianoperatorJr corresponding to differentergen-

values.they must be orthogonal).This sameprocessis then usedto generate
l J . J - 2 ) l J - 1 . J - 2 | 1 a n d ( b y o r t h o g o n a l i tcyo n s t r u c t i o nl )J - 2 . J - 2 ) .
and so on.

(i) The stateswith maximum and minimum M-values

We begin with the statel-I../) havingthe highestM-r'alue.This statemust be
formedby takingthe highestlz andthe highestrr' values(i.e..m : j andnt' :
7'). and is givenby
(2.5
14)

Only this one product is neededbecauseonly the one term with rt - .7 tnd
m' : i'contributesto the sum in the aboveCG series.The state

I r ,- r ) : t i ,- . i | j ' . - j ' ) ( 2 .s15)

with theminimum M-vaiue is alsogivenasa singleproductstate.Noticethatthese
stateshaveM-valuesgivenas +U + /'): sincethis is the maximumM-value,it
must be that the J-value correspondingto this stateis J : j t .i' .

(ii) Stateswith one lower M-r,alue but the same "/-value

Applying J- to I.I. J), and expressingJ- as the sum of lowering operatorsfor
the two individual ansularmomenta:

J_:J_(t)+J,(2) ( 2 .l 5 6 )
gives

J - l J . h : 1 1 1 J ( J+ l ) - J ( J - I ) ) r i r l " r . " l/ )-
: ( J - ( l ) + r - ( 2 ) ) l j ,i ) l i ' j ' )
= h u u + 1 )- j u - l ) ) ' / , | j , - 1 | j " j ' )+ n 1 1 ' 1a. 1 '1
_ j , l j ,_ t ) | , / r U
j t .U , j. , _ r ) . (2.157)
This result expresseslJ, J - l) as follows:

l J . J - 1 ) : [ { , r ( ,+i 1 ) - j ( j - l ) } ' / ' l J j, - t ) l i ' . i ' )

+ I . i ' U+' t ) - i ' ( i ' - 1 ) ] ' / 2i|)i U
, ' ,i ' - 1 ) l
lJ(J + t) - J(J - l)1-rr:. (2.i58)

that is, the IJ, J - I ) state,which hasM : -/ - l, is formedfrom thetwo product

states17,j - lllj' ,7') and lj, j)l j' , j' - 1) thathavethis same M-value.

(iii) Stateswith one lower "I-value

To find the state lJ -1,J -l) that has the same M-value as the one found
abovebut one lower ,/-value,we must constructanothercombinationof the two
p r o d u c t s t a t e s w iMt h: J - I ( i . e . , l j ,j - l ) l j ' , j ' l a n d l j . j ) l j ' . j ' - 1 ) ) t h a t
 A n g u l a rm o m e n t u m
89

is orthogonalto the combinationrepresenting

fJ, J _ t):after doing so,we must
function so it is properry
io.ruri,.a.';;;;r;;r;,
;;;::,ff,:.r"lting the desired
It
r
: t J _ t . J _ t ) : t u u + 1 )_ j ( j _
1)),/21 j )ju. , , j , _ t )
- ti'(j' + t) - j'(j, -
t ) ) t / 2 Uj , _ 1 ) U ,j., ) l
I J ( J + 1 )_ J ( J _ t y _ t r z .
(2.159)
It is straightforwardro showthat this
function is indeedorthogonal
to lJ, J _ t).

(iv) Stateswith evenone lower

,/_yalue
Har-ingexpressedlJ, J _ l) and _
1. J

:::a'Jil;l;;'we
lJ
are
no*
o"oo*J'1ryi:'ffi;[
fi."iJ;l;]
combinati ons",*.; :::T:,',';,1, _il;,L :trt_-rlifr: ir.";,r., ji il
U ,j - 2 ) U ' , j ' ) . I i , i ) t j , . . j , _ 2 ) , a n d t j . j _ 1 ) U , .
j,_
;
arepreciselyas many product "r,
l). Noticetharthere
stateswhose * ,r, values
M-varueas there are totar angular add up to the desired
momentumstatesthat must
(therearethreeofeach in be constructed
this case).
stepsneededro find thestat; -
l-l 2, J -2) areanarogous ro thosetaken
.rJ::

( r ) o n e f i r s r a p p l i e st.o, /l_- r _
l . - r _ l ) a n dt o l J , J _ l ) r o o b r a i n
and1,./ . ./ _ 2). respectively; l J _ 1 .J _
os.ombinotion,,) r l i ' i - 2 ) t i ' ' i ' ) ' U ' i ) U2")i ' - 2 ) .
ti,i- tlti'.i'-,,.
(rr) Onethenconsrructs _
tJ 2. J _ 2) asa linear
ti.i)ti'.i' - 2),aw)
tj.j _t)ti,
./,_r)*"::,1?il:::illny.j,,;,-lli;J,,
fbundfortJ_t._
J2)and
tJ.J_2).

once l-r - 2, J - 2) is
obtained.it is then possibre
lJ,J-3),tJ-I.J-3), to move on to form
r1d lt_z,j_:,1by a p p l y i n -g/ _ t o t h e t h r e e
tnpreceding
apptication
ortn. p.o..r,lonJ,o
;:::yiTi
-,,| "i::n,,1";U',11;i,;ilJ##
I thenrbrm
.i, tj
:: :)li;
obtarne f,*:ll,;;l*j
f odr l J , J - 3 ) . t J _ I , J _ 3 ) , a n d
Againnoticethatther tt _Z.J _3).
rn; ;;'; ;. ;; ""ffi: TH'.'jt'nc orrec
h.r.r"u, t numberorproduct states(rour
states
andrherotal ""*",- momentum r,"r;,;.'.o
:q"u?in.*l#il,j::?:TJ,;
f i"'J::iTH:l#::::r:nsets
d;;; ;' (i t'''tii,"i c;,;;:i,,.,",
1l1nitary;bec-,",,,""0T*Xiliiil?,i,?;I|J"'Jllff
setscontainingthe ::"."T,n:Tffi
:i;
samenumberof functions.
90 M o d e l p r o b l e m st h a t f o r m i m p o r t a n ts t a r t i n gp o i n t s

An example
Let us consideran examplein r.l,hich the spinandorbitalansularmomentaof the
Si atomin its 3Pgroundstatecanbe coupledto producevariousrp.7states.In tl.ris
case,the specificvaluesfor j andj' are.f - "!: I and : L: 1. We could-
.i'
o f c o u r s e . t a k e-. Tl : I a n d . l ' : S : l . b u t t h e f i n aul , a y e f u n c t i o n s o b t a r n e
would spanthe samespaceas thosew.eareaboufto determine.
Thestatewith highestM-valueis the-rp(,/|1s : 1. M t : I ) state.u,hichcanbe
represented by the productof an ncvspin function(representing S : i . M s : 1)
anda 3p13p6spatialfunction(represenring L : l. M t : l), u,heretheIirstfunc-
tlon corresponds to the first open-shellorbitaland the secondfunctionto the
secondopen-shellorbital.Thus. the maximurr M-value is M:2 and corrc-
spondsto a stater"'ith -/ : 2:

lJ :2. 1 4: l) : 1 2 .l \ : d d 3 p . . 1 p , , . ( 2 .l 6 0 )

Clearly.the state12.-2) would be given as bt'J3p_t3po.

The states12,l) and 11.l) u'ith one lowerll{-valueareobtainedby applying
J - : S * I L _ t o i 2 . 2 ) a sf o l l o w s :

J 1 2 . 2 :) f r { 2 ( 3- ) 2 ( l) } r / : 1 2l ).
: ( S -+ l _ _ ) a c r 3 p 1 3 p o . (2.161)
To apply s* or l- to oa 3p13ps.one mustrealizethat eachof theseoperators
is. in turn,a sum ofloweringoperatorsfor eachofthe two open-shellelectrons:

S-: S (l)+S-(2). (2.t62)

L_-L_(t)+L (2). (2.r63)
The resultabovecan thereforebe continuedas

( S - * Z - ) a a 3 p r 3 p:o l i l l l 2 ( 3 1 2 ) - | l 2 ( - 1 1 : ) l 1 t 2p a 3 p , 3 p s
+ t f t / 2 ( 3 1 2-) | / 2 ( - 1 / 2 i l t i 2 u p 3 p l p s
* hll (2) - I (0)|r1:oo3pn3p6
+ h l l ( 2 )- 0 ( - I ) l r / r a a 3 p3rp - , . (2.t64)
So,the function12.l) is givenby (ao3p63psviolatesthe pauli principle.so it is
removed)

12.1) : [do3n'3Ru
* uB3p13p()
r {2}tt2ua3py3o_l
/2. ( 2 r. 6 5 )
which can be rewrittenas

12.l) : l(fla + uB;lpr3po+ l2]tt2os3p'3p_,)

/2. (2.166)
writing the result in this way makesit clear that
12,l) is a combinationof the
p r o d u cst t a t e sl , s: 1 , M s : 0 ) l Z : l , M t : l ) ( t h et e r m sc o n t a i n i n e
ls: l.
 A n g u l a rm o m e n t u m
91

. t / 5= 0 ) : 2 - t i 2 1 a B + F a ) ) a n dl s : 1 , l v l s : l ) l Z : l . l u l t : 0 ) ( t h et e r m s
g : 1 . f u5l : 1 1 - a . ' ) .
c o n t a i n i nl S
T o f o r m t h e o t h e r f u n c t i own i t h M : 1 ,t h e i 1 , l ) s t a t e , w e m u s t f i n d a n o t h e r
o f
c o r n b i n a t i o nl S : 1 .M 5 : 0 ) l I : l , M 1 : 1 ) a n < ll S : l , M 5 : 1 7 1 7 :
l.,l[L :0) that is orthogonal to 12.l) and is normalized.Since

1 2r. ) : 2 - ' ' t l l s: t . i n s : 0 ) i r : t . M 1 - t ) + l " t : l . , r / . r : l )

x lI = l,Mt:0)),
e.167)
we immediatelyseethat the requisitefunction is

l l . l ) : 2 - ' ' : [ l S - I , M r = 0 ) l Z: t . M L : l ) - l S = l , , V s : t )
x lI = l, illr:97)1. (2.r68)
In thespin-orbitalnotationusedabove.this stateis

l l , l ) = f { f l " + " B 1 l p , 3 p-u { 2 } , ' 2 a a 3 p } p - 1 )2 . ( 2 r.6 9 )

f
T h u sf a r ,w e h a v ef b u n d t h e3 p . 7s t a t ew
s i t hJ : 2 , M : 2 : J : 2 . ilt: l: and
l - l i . t - l

T o f i n d t h e 3 Psl r a t e s w i tJh : 2 , M : 0 : J : l . M : 0 ; a n d - I :
0.M : 0.
we must once agarnapply the -/_ tool. In particular,u,e applv,/_
to 12.l) to
o b t a i n1 2 . 0 )a n d u , ea p p l y- / _ t o l l . l ) t o o b t a i n
l l . 0 ) , e a c ho f w h i c h w i l l b e
e x p r e s s ei ndt e r m so f l . S : l . M 5 : 0 ) l l : l . M 1 ,: 0 ) .
l.t : l. M.;: l)lt :
l . i l [ 1 : - l ) . a n dl S : I . M 5 - _ l ) l L : l . M 1 : i ) . T h e
1 0 . 0 )s t a t ei s t h e n
constructed to be a combinationof thesesameproductstateswhich is orthosonal
t o 1 2 . 0 )a n dt o l l . 0 ) . T h e r e s u l t sa r ea s f o l l o w s :

) . J : 2 .t r : 0 ) : 6 " . 1 2 1 l . 0 ) l t . 0I l) .+l ) l l ._ l ) + I i . _ l ) l t . r ) 1 .( 2 . r 7 0 )
l J : r . , v : 0 ) : 2 - , , , | 1 .l ) l l ._ l ) _ l l . _ l ) l l . l ) 1 .
Q.t7t)
l . J: 0 . , V : 0 ) : 3 , , [ l r . 0 ) l t . 0 ) _ l ll ). l l . _ l )_ l t . _ t ) l l ,1 ) 1 . ( 2 . 1 7 2 )
where. in all cases.a shorthandnotation has
been used in which the
l-t. ,I'l.s)|L . llt L) productsstatedhavebeenrepresented by their quantumnumbers
with the spin function arrvaysappearingfirst in the product.
To finally express
all threeof thesenervfunctionsin termsof spin-orbitalproducts
it is necessary
to gi'e the l.l. rvy) lL . l,l t.) procluctswith M : 0 in terms
of theseproducts.For
thespin functions,we have

lS: l. Ms = l) = ao. (2.173)

l S : I , M s - 0 ) = 2 - t t 2 ( c t +p p a ) ( 2 . 117)
s -l) : ff .
l S : I , r l . 1= (2.t1s)
For the orbital product function, we have

It : i. Mt = I):3pr3p,,. (2.176)
JZ M o d e lp r o b l e m st h a t f o r m i m p o r t a n ts t a r t i n gp o i n t s

iZ : l . A 4 r : g y = 2 1 , ' : ( 3 p t , 3* p3op ; 3 p , 1 . (2.177)

lL:1,Mt=-l) :3po3p_i. (2.178\

Coupling angular momenta of equivalent electrons

If equivalentangularmomentaare coupled(e.g."to couplethe orbital
angular
momentaof a p2 or d3 configuration), one mustusethe follorving.,box'.method
to determinewhich of the term symbols'iolate the pauli principle.To carry
out
this step.one forms all possibleunique(determinental)
productstateswitli non-
negativeM 1 andM5 valuesand arrangesthen.rinto groupsaccordingto
rheirM 7,
and M5 r'alues.For example.the boxesappropriateto the p: orbital occupancy
are shownbelow:

M v 2

M, I lptcttpoctl Iprap.-ral
U lppptlll l p t a p o f l . l p n a p t f l l p ' c l p _ t f l l . l pr a p r f J ll.p o c r p o f J l

There is no need to form the correspondin-qstates r.l,ithnegative M1,

ornegative
Ms values because they are simply "mirror images" of those listed
above. For
e x a m p l e ,t h e s t a t e w i t h M 1 : -l and Ms: -l is
l p _ r f j p o f J l .w h i c h c a n b e
o b t a i n e df r o m t h e M L : I. Ms: I s t a t e l p l n p e a l b y r e p l a c i n go b y p a n d
replacingpr by p_r.
Given the box entries, one can identify those term symbols that arise
by appry-
ing the following procedure over and over until all entries have
been accounted
for:

(i) one identifiesthe highestM5 value (this givesa varueof the total spin quantum
numberthat arises,s) in the box. For the aboveexample.the answeris :
s l.
(ii) For all productstatesof /?,.rM5 value,one identifiesthe
highestM7_value(this
givesa valueof the total orbital angularmomentum,Z, that can arisey'rr
llrs .f).
Forthe aboveexample,the highestM7 withinthe M5 = I statesis M :
t | (not
ML :2), henceI : l.
(iii) Knowing an ,s, Z combination, one knows the first term
symbol that arisesfrom
this configuration.In the p2 example,this is ip.
(iv) Becausethe level with thesez and s quantumnumbers
conrains(2L + 1x2s + l)
stateswith M1 and M5 quantum numbers running from _Z to Z and from _.!
to
.1,respectively, one must removefrom the originarbox this numberof product
states.To do so, one simply erasesfrom the box one entry u,ith eachsuch
ML and
M5 value.Actually,sincethe box needonly show thoseentrieswith non-necatrve
M7 andM5 values,only theseentriesneedbe explicitly deleted.In the 3p
example,this amountsto deletingnine productstateswith ML, Ms values
of
I , l ; 1 , 0 ;l , - l ; 0 , 1 ; 0 , 0 ; 0-,t ; _ 1 . 1 ;_ 1 , 0 ; _ 1 . _ 1 .
(v) After deletingtheseentries,one returnsto
step 1 and carriesout the processagaln.
Forthe p2 example,the box after deretingthe first nine productstates
rooksas
 Angular momentum VJ

follo*'s (thosethat appearin italics shouldbe viervedas alreadycancelledin

countingall ofthe lP states):

,llt 2

,!1, I lpraprpl lptap-pl

0 lpraprdl lpspofJl.lpnap'dl lpsp;Fl,lp-roprEl, lpoapodl

It shouldbe emphasizedthat the processof deleting or crossingoff entriesin

variousML, lvls boxes involves only counting how many statesthere are; by
no meansdo we identify the particularL . S, M t. M5 wave functionswhen we
crossout any particularentry in a box. For example,when the lprcvpodlprod-
uct is deletedfrom the Mt: I. r!1s:0 box in accountingfor the statesin
3P level. we do not claim that lP
the lpgpofrl itselfis a memberof the level;the
lpocvp rf I productstatecould just as well be eliminatedwhen accountingfor the
lP states.
Returningto the p2 exampleat hand.after the lP term symbol'sstateshave
beenaccountedfor, the highestM5 valueis 0 (hencethereis an ^S: 0 state),and
within this Me value,the highestII 1,valueis 2 (hencethereis an L :2 state).
T h i s m e a n st h e r ei s a r D l e v e lr , v i t hf i v e s t a t e sh a v i n gM t : 2 . 1 . 0 . -1. -2.
Deletingfive appropriateentriesfiom the abovebox (again denotingdeletions
by italics)leavesthe followingbox:

,llt 2

M, l Ip l a p x u l lptap tul
U lpsptlJl l : r a p r t l J l . ' , p r u p r f l lp t u p r f 1 . \ p _ t a p t f l l . l p , r a p o f l

The only r:mainingentry,which thus has the highestM5 and M1 values.has

t/s : 0 andM y : 0. Thusthereis alsoa lS levelin the p2 configuration.
T h u s ,u n l i k et h en o n - e q u i v a l e2npt r 3 p i c a s e i,n w h i c hl B ' P ,l D , ' D , 3 S , a n d
lS levelsarise.
only the lP. lD. andls arisein the p2 situation.It is necessary
to carry out this "box method" wheneverone is dealingwith equivalentangular
momenta.
If one has mixed equivalentand non-equivalent angularmomenta,one can
determineall possibtecouplingsof the equivalentangularmomentausingthis
methodandthenusethesimplervectorcoupling methodto addthenon-equivalent
angularmomentato each of thesecoupledangularmomenta.For example.the
p:dl configurationcan be handledby vectorcoupling(usingthe straightforwarcl
non-equivalent procedure) Z : 2 (thed orbital)and,S: I l2 (thethird electron's
spin)ro eachof 3p,rD,and iS. The resultis .F, .D, .p, 2R 2D,,q ,G, 2F.2D,2p
2 S .a n d2 D .
94 M o d e lp r o b l e m st h a t f o r m i m p o r t a n ts t a r t i n gp o i n t s

2.8 Rotationsof molecules

2 . 8 . 1R o t a t i o n aml o t i o nf o r r i g i dd i a t o m i ca n d l i n e a r
p o l y a t o m i cm o l e c u l e s
This schrodinger equation relatesto the rotation of diatomic and linear
polyatomic molecules.lt also ariseswhen treating the angular motions
of electronsin any sphericallysymmetric potential.

A diatomicmoleculervith fixedbond lengthR rotatingin tlreabsenccof any

externalpotentialis describedby the followingSchrodinger
equation:
-h'l I a (ri,,r1')--f-!-)r:ur
( 2 . 17 9 )
2/t LR: s"*/ rl', \ r J H/ R- sin- F i)Qt )

t : ,'t ,
-- 'F' ,"t ', (2.180)
2lLR:

wherez- is the squareof the total angularrnomentumoperatort,? + tl

+ tl
expressed in polarcoordinates
above.The anglesd and@describetheorientatlon
of the diatornicmolecule'saxis relativeto a laboratory-fixed
coordinatesysrem.
andpr is the reducedmassof the diatomicmoleculeU : il.t1ttt2f(mr * ri:). The
differentialoperatorscan be seento be exactlythe sameas those that arosern
the hydrogen-likeatom caseas discussedabove.Therefore,the samespherical
harmonicsthat servedas the angularpartsofthe wavefunction in the hvdrogen-
atomcasenow serveasthe entirewavefunctionfor the so-calledrigid rotor: :
v/
YL.u@.@).Theseare exactlythe samefunctionsas we plottedearlierwhen we
graphedthes (Z : 0), p (L : 1),andd (L : 2) orbitals.Theenergl,eigenvalues
correspondingto eachsucheigenfunctionare given as

_ h2JtJ+ll
tL: :BJ\J-ll (2.181)
zttn,
andareindependentof M. Thus eachenergylevelis labeledby J andis (2J + 1)-
foid degenerate (becauseM rangesfrom -J to -r). Again, this is just like we saw
when we lookedat the hydrogenorbitals;the p orbitalsarethree-folddegenerate
and the d orbitals are five-fold degenerate.The so-calledrotational consranr
B (definedas tl2121t"R21 dependson the molecule'sbond length and reduced
mass.Spacingsbetweensuccessive rotationallevels(which are of spectroscopic
relevancebecause,as shown in chapter 5, angularmomentum selectionrules
often restrictthe changesLJ ]n J that can occurupon photonabsorptionto 1,0,
and -l) are given by

LE : 311+ 1X-/+ 2)- BJ(J* 1): 2p11* t, (2.182)

Theseenergy spacingsare ofrelevance to microwavespectroscopywhich probes
the rotationalenergylevelsof molecules.In fact,microwavespectroscopy
offers
 R o t a t i o n so f m o l e c u l e s 95

themostdirectway to determinemolecularrotationalconstantsandhencemolec-
ularbondlengths.
The rigid rotor providesthe most commonly employedapproximationto the
rotationalenergiesand wave functionsof linearmolecules.As presentedabove,
themodelrestrictsthebond lengthto be fixed.vibrational motion of the molecule
givesrise to changesin R which are then reflectedin changesin the rotational
energylevels.The couplingbetweenrotationaland vibrationalmotion givesrise
to rotationalB constantsthat dependon vibrationalstateas well as dynamical
couplings.calledcentrifu-eal distortions,thatcausethe total ro-vibrationalenergy
of the moleculeto dependon rotationaland vibrationalquantumnumbersin a
non-separable manner.
"rigid rotor"
Within this model, the absorptionspectrumof a rigid diatomic
moleculeshoulddisplaya seriesofpeaks.eachofwhich corresponds to a specific
J - J f I transition.The energiesat which these peaks occur should grow
linearlywith -r. An exampleof such a progressionof rotationallines is shown
inFig.2.23. The energiesat which the rotationaltransitionsoccurappearto fit
the AE:28(J * l) formula ratherwell. The intensitiesof transitionsfrom
level-/ to level -/ + 1 \,ary stronglywith J primarily becausethe populationof
moleculesin the absorbinglevel varieswith -/. Thesepopulationspr are given.
whenthe systemis at equilibrium at temperaturer, in terms of the degeneracy
(2J + 1)of the-/th levelandtheenergyof thislevelB J(J + l ) by theBoltzmann
formula:

p . r: e - t e J * t ) e x p ( - 8 , / ( , 1+ 1 ) l k T ) (2.r83)
wherep is the rotationalpartition function:

O =ltzL * l ) e x p ( - B - r (+J t ) l k T ) (2.18.1)

For low valuesof -/, the degeneracyis low and the exp(-B-r(J +l)l kr)
factor is near unity. As -/ increases,the degeracygrows linearly but the

Typical
rotationalabsorption
profileshowing
intensityvs. J valueof
the absorbinglevel.
96 M o d e l p r o b l e m st h a t f o r m i m p o r t a n ts t a r t i n gp o i n t s

exp(- B J ( J + 1) I k T ) factor decreases more rapidly.As a result,thereis a value

o f J , g i v e nb y t a k i n gt h e d e r i v a t i v oe f ( 2 J - l 1 ) e x p ( - B J ( J + 1 ) l k T ) u ' i t h r e -
spectto J and settingit equalto zero.

2 J - u ^* l : 1 , 5 f f i . (2.1g5)

at which the intensityofthe rotationaltransitionis expectedto reachits maxrmurn.

This behavioris clearly displayedin the abovefigure.
Theeigenfunctions belongingto theseenergylevelsarethesphericalharmon-
ics Yt.v@, @)which arenormalizedaccordingto

u,@.0)sin0dedQT:3tt,6r.u'. (2.186)

As notedabove,thesefunctionsare identicalto thosethat appearin the solu-

tion of the angularpart of hydrogen-likeatoms.The aboveenergl' levels and
eigenfunctions also apply to the rotationof rigid linearpolyatomicmolecules;
the only differenceis that the moment of inertia / enteringinto the rotatior.ral
energyexpression is givenby

I:lm,R; ( 2 r. 8 7 )

wherenr,,is the n.rass

of theath atomand Ro is its distancefrom the centerof mass
of the molecule.This momentof inertiareplacespR2 in the earlierrotational
energylevelexpressions.

2 . 8 . 2R o t a t i o n am
l o t i o n so f r i g i d n o n - l i n e a m
r olecules
Therotationalkineticenergy
The rotationalkinetic energyoperatorfor a rigid polyatomicmoleculeis

H,o,: Jj /2t" + J: l2tb + J: /2t,., (2.188)

where the lp(k: a,b.c) are the three principal momentsof inertia of the
molecule(the eigenvaluesof the moment of inertia tensor).This tensorhas el-
ementsin a Cartesiancoordinatesystem(K,K' : X,Y. Z), whoseorigin is
locatedat the centerof massof the molecule,that can be computedas

/r.": I m , ( n ' -, n ' r . ) (forK: K') ( 2 . 81 e )

Ir.r,: -LmiR*tRr,.i g o rK I K'). (2.190)

As discussedin more detail in Chapter6, the componentsof the quanturnme-

chanicalangularmomentumoperatorsalong the threeprincipal axesare

J"- -ttcos
x(*,rr\ - ##) - iisinx (2.r9r)
$
 R o t a t i o n so f m o l e c u l e s 97

(."tea-
Jt= ittsinx
# il. ihco'xfi (2.1e21

,
Jc:
- a
-lll:-
( 2 .r 9 3 )
dx

The angles0 . Q, and x are the Euler anglesneededto specify the orientationof
the rigid moleculerelativeto a laboratory-fixedcoordinatesystem.The corre-
spondingsquareof the total angularmomentumoperatorJ2 canbe obtainedas

J2:Jj+t]+Lj
a2 1 /a2
: - ' d 0 : ,- c o t^?a- * *
a2-
2cos0 =
6: \
t2'1941
,i,r, \ro, ax: ).
and the componentalong the lab-fixedZ-axis J7 is -ili\lD6 as we saw much
earlierin this text.

The eigenfunctions and eigenvalues for specialcases

(i) Sphericaltops
when the threeprincipal momentof inertiavaluesare identical.the moleculeis
termeda sphericaltop. In this case,the total rotationalenergycan be expressed
in termsof the total angularmomentumoperatorJ2

H , , ,= J ) l 2 [ . (2.19s)
As a restrlt.the eigenfunctionsof Hro,are those of J2 and Ju as well as J7,
both of which commute with J2 and,with one another.,,/7 is the component
of J alongthe lab-fixedZ-axis and commuteswith Jo becauseJz: -ihAlAQ
and J,, - -iliD/Dx act on different angles.The energiesassociatedwith such
eigenfunctions are

E(J. K. M):h2J(J + I)/2t2 (2.t96\

for all K (i.e., Jn quantumnumbers)rangingfrom --l to -/ in unit stepsand for
all l,I (i.e., Jz quantumnumbers)rangingfrom -J to -/. Each energy level is
therefore(2.1+ l)2 degenerate becausethereare 2J + I possibleK valuesand
2J + | possibleM valuesfor eachJ.
The eigenfunctionsj-l. M. K) of J2, J7 andJo , aregiven in terms of the set
of so-calledrotationmatricesDt.tr.x..

D I - Ll
, J . 1 4K
. \: D ) v K @ . Q x. t . (2.re7)
v il
which obey

J l ) J . ] , t ,K l : h 2L ( L + l ) l J , M , K ) , (2.r98)
J , , I JM, . K ) : h K l J ,M . K ) , (2.t99)
J z l J .M . K ) : t 1 M l JM
, , Kt. (2.200)
oa M o d e l p r o b l e m st h a t f o r m i m p o r t a n ts t a r t i n gp o i n t s

TheseD.7..v.1'fu'ctions areproportionalto the sphericalhar'ronicsYt.tr().0)

multipliedby exp(iKX)"lr'hichreffectsits X-dependence.

(ii) S1'mmetrictops
Moleculesfor which two of the threeprincipalmomentsof incrtiaareequalare
calledsymrnetrictops.Thosefor which the uniquemomentof jnertiais smaller
than the othertwo aretermedprolatesymmetrictops:if the uniquemomentof
inertiais largerthan the others,the moleculeis an oblatesvrnmetrictop. An
An.rericanfootball is prolate,and a frisbeeis oblate.
Again.tlrerotationalkineticenergy.which is the full rotationalHamilronran"
canbe u'ritten in termsof the total rotationalan_tularmomentunroperatorJl and
the componentof angularmomentumalongthe axis with the uniqueprincipal
n'roment of inertia:

H , o , : . / :l 2 l + J : l l l 2 l " - l l 2 l l f o r p r o l a treo p s . ( 2 . 2 0) 1
Hu, : J2l2l + J,2l l2l, - | l2t I for oblatetops. (2.2021

Here,the momentof inertia ,1denotesthat momentthat is conrmonto trvo direc-

tions;thatis, / is thenon-unique
momentof inertia.As a result.theeigenfunctions
of H,o, are thoseof J: and Ju or J, (and of J7), andthe correspondingenergl,
levelsare

E ( J ,K . M ) = h 2J ( J + j ) 1 2 t 1+ f f 1 2 1pt r , , _ t l 2 t ) (2.2031

for prolatetops,

E ( J .K . M \ : t r 2 J ( J + t ) / 2 r 2+ t t 2 K 2 l t / 2 t-, l l 2 r ) (2.20-+)
for oblate tops. again for K and M (i.e., Jo or J, and J7 quantumnumbers,
respectively)rangin-qfrom -J to J in unit steps.Sincethe energynow depends
on K, these levels are only 2J * 1 degeneratedue to the 2J * 1 different M
valuesthat arise for eachJ value. Notice that for prolate tops, because1., is
smallerthan I, the energiesincreasewith increasingK for given -/. In contrast.
for obiatetops, since ,I..is largerthan 1, the energiesdecreasewith K for given
J. The eigenfunctionslJ. M , K) arethe samerotationmatrix functionsas arise
for the spherical-topcase,so they do not requireany further discussionat this
time.

(iii) Asymmetric tops

The rotationaleigenfunctionsand energylevelsof a moleculefor which all three
principal momentsof inertia are distinct (a so-calledasyrnmetrictop) can not
easily be expressedin terms of the angularmomentumeigenstatesand the J,
M, and K quanfumnumbers.In fact, no one has ever solvedthe corresponding
Schrodingerequationfor this case.However,given the threeprincipal moments
ofinertia Io, I6,and I",amatrixrepresentationofeach ofthe threecontributions
 V i b r a t i o n so f m o l e c u l e s qq

to the rotationalHamiltonian
12 r2 12
r.; J' J:
u - '
- ( 2 . 2 0)5
2t, 2tr, 2t.
canbe formedwithin a basissetof the { lJ. M . K)l rotation-matrix
functionsdis-
cussedearlier.This matrixwill not be diagonalbecausethelJ. M . K) functions
are not eigenfunctionsof the asymmetrictop H.o1.However,the matrix can be
formed in this basisand subsequentlybrought to diagonalform by finding its
eigenvectors{Cn..r.v.x}andits eigenvalues {f,,}. The vectorcoefficients
express
the asymmetrictop eigenstates as

v ' @ ' o''' = / a C n' t ' K l JM

' 'K ) ' (2.206)

Becausethe total angular momentum /2 still commuteswith H,o,.each such

rvill containonly one -/-value.and hencev,, can also be labeledby a
eigenstate
./ quantumnumber:
\ U ,. t @Q
. ,X ) : C , ,r \ rK ) J .M . K ) . (2.207)
I

To form theonly non-zeromatrixelementsof H,o,within the lJ. M, K) basis.

one can use the fbllowing propertiesof the rotation-matrixfunctions(see.tbr
example.R. N. Zare.AngularMomennun,JohnWiley.New york (198g)):

( J , M . K i J , i l . t i. l t . K ) : \ J . M . K I J ; . J . M , K )
. V l * J : i J . L r .K )
: t l 2 \ J . , r . tK
:h:[J(,]+l)-K:1. (2 . 2 0)8

(2.209)

( J . , V . K l JJ ,. it t . K + 2 ) = - . ( J . , V . K l J ; l J . M+. 2K1
= h ' 1 . / (+J l ) - K ( K + l ) l r: u ( - /+ l )
- (( + lxK + 2)lr':, (2.210)

(2.2rl)
Eachof theelements of .1 . J,i.andJolmust,of course,bemurtipriedrespectivelv.
by 1121,.| 121,,.and I 12l6 and summedtogetherto form the matrixrepresenra-
ttonof 11,u,.
The diagonalization of this matrixthenprovidesthe asymmetrictop
energlesand rvavefunctions.

2.9 Vibrationsof molecules

T h i s S c h r o d i n g e er q u a t i o nf o r m s t h e b a s i sf o r o u r t h i n k i n ga b o u t b o n d
stretchingand angle bending vibrations as well as collectivevibratioi-rs
calledphonons in solids.
100 M o d e lp r o b l e m st h a t f o r m i m p o r t a n ts t a r t i n gp o i n t s

Theradialmotionof a diatomicmoleculein its lorvest(J : 0) rotationallevel

can be describedby the following Schrcidinger
equation:

tl.lllr

l r ' h e r ep i s t h e r e d u c e dm a s sF : m 1 m 2(f n t 1 + m 2 ) o f t h e t w o a t o m s .I f t h e
moleculeis rotating.thenthe aboveSchrodinger equationhasan additionalterm
J ( J + 1 ) h 2 / 2 L L r2 l r o n i t s l e f t - h a n d s i d e . T h u s , e a c h r o r a t i o n a l s t a t e ( l a b e l e
the rotationalquantumnumberJ) hasits own vibrationalSchrodinger equation
and thus its own setof vibrationalenergylevelsand u'avefunctions.It is com-
mon to examinethe J : 0 vibrationalproblemand then to usethe vibrational
levelsof this stateas approximations to the vibrationallevelsof srateswith non-
zero J values(treatingthe vibration-rotationcoupling via perturbationtheorv
introducedin Section4.1).Let us thusfocuson the J : 0 situation.
By substitutinglr : F(r)lr into this equation,one obtainsan equationfor
F(r) in which the differentialoperatorsappearto be lesscomplicated:

li1 d2 F
-i * I/(r\F: EF (1.213)
a,-,
This equationis exactlythe sameas the equationseenearlierin this text for the
radial motion of the electronin the hydrogen-likeatomsexceprthat the reduced
mass/.. replacesthe electronmass/ri and the potentialI/(r) is not the coulomb
potential.
Ifthe vibrationalpotentialis approximatedas a quadraticfunctionofthe bond
displacementx : t' - r. expandedabout the equilibrium bond lengthr. where
Z hasits minimum.

V=1/2k(r-r,)2 (2.214)

the resultingharmonic-oscillatorequationcan be solved exactly.Becausethe

potential Iu grows without bound as x approachesoc or -oo, only bound-state
solutionsexist for this model probiem. That is, the motion is confinedby the
nature of the potential, so no continuum statesexist in which the two aroms
boundtogetherby the potentialare dissociatedinto two separareatoms.
In solvingtheradialdifferentialequationfor thispotenrial,thelarge-rbehavior
is first examined.For larger, the equationreads

#:)r.'(# ), (2.2ts)

where x : r - re is the bond displacementaway from equilibrium. Defining

q:gtklh2)t/4x as a new scaledradial coordinateallowsthe solution
of the
large-r equationto be written as

fiu.g.": exp(-(2 l2). (2.2t6)

 V i b r a t i o n so f m o l e c u l e s
101

The ge'eral solution to the radial equationis then expressedas this large-r
solutionmultiplied by a powerseriesin the ( variable:

F : :- er v! n \/ -/ . tLl / _? t \ \r - n erl-
- r , /
(2.217)
n=v

wherethe c, are coefficientsto be determined.Substitutingthis expressioninto

the full radial equationgeneratesa set of recursionequationsfor the c,, ampli_
tudes.As in the solutionofthe hydrogen-likeradial equation,the seriesdescribed
by thesecoefficientsis divergentunlessthe energy E happensto equarspecific
values.It is this requirementthat the wavefunction not divergeso it can be nor_
malizedthat vields energyquantization.The energiesof the statesthat anseare
gir en by

E,, : h(.k
/ p)t'2(n + | /b. ( 2 . r28 )
andthe ei-eenfunctions
aregivenin termsof the so-calledHermitepolvnomiars
H,,(v)as lbllows:

t t 2 1 a / ry r r a
Ir,(x) : qn!2") e x p l - a - r : / 2 ) H , ,( u t i ) x ) , ( 2 .r2e )
wherea : (ktrtl/1-1ti2.
within this harmonicapproximation
to the potential.the
vibrationalenergylevelsarc'evenlyspaced:

A E : E , + r - E , : t t , k lp ) t ' 2 (2.220\
In experimental datasuchevenlvspacedener,syIevelpatternsare seldomseen:
most commonll,.one finds spacingSEr*l _ 8,, that decreaseas the quantum
numbern increases. In suchcases.one saysthat the progression of vibrational
l e v e l sd i s p i r ls u n h a r m o n i c i r v .
Because theHermitefunctionsH,,areoddorevenfunctionsof.r (depending
on
whethern is odd or even).thervavefunctionsry',, (-r) areodd or even.This splittin_q
of thesolutionsinto tr'",o distinctclasses is an exampleof theeffectof symmetry;
tn this case.the symmetryis causedby the symmetryof the
harmonicpotential
rvith respecrto reflectionthroughthe origin along the -r-axis(i.e..
chansrn,s .r-
to --r). Throughoutthis text. many symmetriesarise:in
eachcase.symmerry
propertiesofthe potentialcausethe solutionsofthe
Schrtidingerequationto be
decomposed into varioussymmetrygroupings.Suchsymmetrydecompositions
areofgreatusebecause theypror,ideadditionalquantumnumbers(i.e..symmerry
labels)by which the wavetunctionsandenergiescan be labeled.
The basic idea u'derlying how such symmetriessprit the solutions
of the
Schrcidinger equationinto diftbrent classesrelatesto the fact that a symmerrv
operator(e.g..the reflectionplane in the aboveexample)
commuteswith the
Hamiltonian.That is. the symmetryoperatorS obeys

SHq/ : HS:\y. (2.22r)

 102 M o d e lp r o b l e m st h a t f o r m i m p o r t a n ts t a r t i n gp o i n t s

So -l leavesH uncl.ranged as it actson ff (this ailows us to passs throughff in

the abo'e equation).Any operatorthatreavesthe Hamirtonian
(i.e..the energv)
unchanged is calleda syrnnetrvoperator.
If you ha'e never learnedabout horv point group
symmetry can be usedto
helpsimplifythe solutionof the Schrcidi.ger equation.thiswouidbe a goodtime
to lnterruptyour readingand go to chapter4 andrea<i
the rnaterialthere,
The harmonicoscillatorenergiesand 'u'avefunctions
conrprisethe simprest
reasonable model for 'ibrational nlotion.vibrationsof a polyatomic
molecule
areoftencharacterized in termsof incliridualbo'd-stretchi'ganda'gle_bending
motions.eachof which is. in rurn.approximated ha'.ronicalry.This resurtsin a
total vibrationalwavefunctiontl.ratis writtenas a product
of funcrions" one fbr
eachof the vibrationalcoordinates.
T\l'oof the mostseverelimitationsof the harrnonic
oscillatornrodel,the lack
of anharmonicity(i.e., non-uniforn.r ener-qvrevel spacings)and lack of. bond
dissociation. result from the quadraticnatureof its potenrial.By
introducing
modelpotentialsthatallow for properbonddissociation (i.e..thatcio,ot increase
without boundas,t --r co). the nrajorshortconrings
of the harm'nic oscillator
picturecan be overcome.The so-calledMorsepotentiar
(seeFis.2.24)

I/(r) - D"ll - exp[-a(r -.")]]: (2.222\

is often usedin this regard.
In the Morsepotentialfunction.D" is the bond dissociation
energy,r. is the
equilibriumbond length, and a is a constantthat
the ,.steepness,,
characterizes
of the potentialand thus affectsthe vibrational frequencies.
The advantageof.
usingthe Morsepotentialto improveupon harmonic_oscillator_level
predictions

Morse
potentialenergyas a -6
function of bond length. 2
Internuclear
distance
 V i b r a t i o n so f m o l e c u l e s 103

is that its energyler''elsand wave functionsare also knorvnexactly.The energres

aregiven in termsof the parametersof the potentialas follows:

u,=,(x),[ft.;)
(,+ 1)'o(X)'''
4,n
. (2.223)

wherethe forceconstantis givenby k :2D, a2. The Morsepotentialsupporrs

bothboundstates(thoselying belowthedissociationthresholdfor which vibration
is confined by an outer turning point) and continuum stateslying abovethe
dissociationthreshold.Its degreeofanharmonicityis governedby the ratio ofthe
harmonicenergy/1(klp;tlz to the dissociationenergyD..
The eigenfunctionsof the harmonicand Morsepotentialsdisplaynodalchar-
acteranalogousto what we haveseenearlierin the particle-in-a-boxmodelprob-
lems.Namely, as the energy of the vibrational state increases,the number of
nodesin the vibrationalwave function also increases.The state having vibra-
tional quantumnumberv has v nodes.I hope that by now the studentis getting
usedto seeingthe numberof nodesincreaseas the quanfumnumberand hence
theenergygrows.