C h a p t e r1

The basicsof quantum mechanics

1.1 Why quantum mechanicsis necessaryfor describing

molecularproperties
we krow that all molcculesare made of atoms which. in turn. containnu-
clei and [Link] I discussin this introcjuctory
section,the equationsthat
govern the motions of electronsand of nuclei are not the familiar Newton
equatrons.

F:ma (l.l)

but a nervset of [Link] scientistsfirst

studiedthe behaviorof electronsand [Link] tried to interprettherr ex-
perimentalfindin-ssin termsof [Link] suchatrempts
[Link] in a waythat
simply is not consistent rvith the Neu'tonequations. Let me norvillustratesorne
ofthe experimentaldatathat gaverise to theseparadoxesand shorvyou how the
scientists ofthoseearlytimesthenusedthesedataro suggestnew equatrons that
theseparticlesrnight obcy.I want to stressthat the Schrcidinger equationwas not
derivedbut postulated by thesescientists. In fact,to date,rlo one hasbeenable
t o d e r i r et h e S c h r c i d i n geeqr u a t i o n .
[Link] pioneeringwork of Braggon ditrractionof x-raysfiom planesof
atomsor ions in crvstals,it was knownthat peaksin the intensityof ditliacted
x-rayshavin,u wavelength i rvouldoccurat scattering anglesg determinedby the
larnousBragg equation:

nt : 24 tinp (1.2)

where d is the spacingbetweenneighborin,u planesof atomsor ions. These

quantitiesare illustratedin Fig. I . I . Therearemany suchdiffractionpeaks,each
l a b e l e d b y a d i f f e r e n t v a l utehoefi n r e g en r( n - 1 , 2 . 3 , . . . ) . T h e B r a - s - q f o r m u l a
can be derivedby consideringwhen two photons,one scatteringfrom the second
plane in the figure and the secondscatteringfrom the third plane,will undergo
[Link] condition is met when the ,.extrapath leneth,'
 T h e b a s i c so f o u a n t u mm e c h a n i c s

Scatteringof
t w o b e a m sa t a n g l e6
f r o m t w o p l a n e si n a
crystalspacedby d.

coveredby the secondphoton (i.e.,the lengthfront pointsA to B to C; is an

integermultipleof the wavelengthof the photons.
The importance of thesex-rayscatterin_e experiments to the studyof electrons
and nuclei appearsin [Link] Davissonand [Link] 1927, u'ho
scatteredelectronsof(reasonably) fixedkineticenerg)'E from metalliccrvstals.
Theseworkersfoundthatplotsofthenumberofscattered electrons asa functionof
scattering
angled displayed "peaks"
atangles6 thatobeyeda Bragg-likeequation.
The startlingthing aboutthis observation is that electronsare [Link] the
Bragg equationis basedon the propertiesof [Link] important observation
[Link] Davisson-Germerexperimentswas that the scatteringangles6
observedfor electronsof kinetic [Link] couldbe fit to the Braggni, : 2d sin0
equationif a wavelengthwere ascribedto theseelectronsthat u'asdefinedby

)":hl(2m"E)t'2. (1.3)

where rl. is the massof the eiectronand i is the constantintroducedby Max

Planckand Albert Einsteinin the eariy 1900sto relatea photon'senergv,Eto
its frequencyy I'ia .E : /rt,. Theseamazingfindingswere amongthe earliestto
suggestthat electrons,u,hich had alwaysbeen viewed as particles,might have
somepropertiesusually ascribedto [Link] is. as de Broglie suggestedin
1925,an electronseemsto havea wavelengthinverselyrelatedto its momenfum,
andto displaywave-typediffraction.I shouldmentionthat analogousdiffraction
u'asalsoobservedwhenothersmalllight particles(e.g.,[Link],nuclei,
and small atomic ions) were scatteredfrom crystal [Link] all such cases,
Bragg-likediffractionis observedand the Bragg equationis found to governthe
scatteringanglesifone assignsa wavelengthto the scatteringparticleaccordingto

),:hlQmE)1t2, (1.4)

where rl is the mass of the scatteredparticle and /r is Planck's consranr

( 6 . 6 2x l 0 - 2 7e r gs ) .
 W h y q u a n t u mm e c h a n i c si s n e c e s s a rfyo r d e s c r i b i n gm o l e c u l a rp r o p e r t i e s

l,.inm

- c
a h
c a
ot c
c

I VisibleI

*.'{fllllrilt-Irffi
frlllill ""'-*,,-onllrll spectrumof atomic
Emission

Anarvsis Paschen h y d r o g e nw i t h s o m e
lrrn lines repeatedbelow to
illustrate t h e s e r i e st o
Llf Brackett w h i c ht h e y b e l o n g .

Theobservation thatelectronsandothersmalllight particlesdisplayivave-like

behavior [Link]
important because theseparticlesarewhatall atomsandmolecules
are made of. So, if we want to fully understandthe motions and behaviorof
[Link] surethat*e can adequately describesuchpropertiesfor
theirconstitr"rents.
Because the classical
Newtonecluations do not containtactors
that sr-rggcst
wave propertiesfor electronsor nuclei mo'",ingfreely in [Link]
abovebehaviorspresented significantchallenges.
Anotherproblemthat arosein earlystudiesof atomsand moleculesresulted
fiom the stLrdyof the photonsemittedfrom atomsand ions that had beenheated
or [Link](e.g.,by electricdischarge).It was found that eachkind
of atom (i.e.,H or C or O) ernittedphotonsrvhosefrequencies u wereof very
[Link] exampleof suchemissionspectrais shownin Fig. I .2
fbr [Link] the top panel,we seeall of the linesemittedwith thcir
wavelengths indicatedin nanometers. The other panelsshow horvtheselines
havebeenanalyzed(by scientists rvhosenarnesareassociated) into patternsthat
relateto the specilicenergylevelsbetweenwhich transitionsoccurto emit the
corresponding photons.
In theearlyattemptsto rationalizesuchspectrain termsof electronicmotlous.
one describedan electronas rnovinsaboutthe atomicnuclei in circularorbits
such as [Link] in Fig. 1.3.A circularorbit was thoughtto be stablewhen the
outwardcentrifugalfbrcecharacterized by radiusr. and speedu (rr.u2/r) on the
electronperf-ectlir
counterbalanced theinwardattractit,eCoulombforce(Ze2 l121
exertedby the nucleusof chargeZ:

-
n.r- ,'t' = Ze-lt . (1 . 5 )

This equation,in turn, allows one to relatethe kinetic energy [Link] to the
CoulornbicenergyZe2lr, and thusto expressthe t o t a le n e r g yE o f a n o r b i t i n
 T h e b a s i c so f q u a n t u mm e c h a n i c s

terrnsof the radiusof the orbit:

I -l
E : - r t t . . r -- Z t ' - , r ' : - Z L ' -/ t ' . (1.6)
f ' ' 1

The energycharacterizingan orbit ofradius r. relatil'eto the 6 : 0 reference

of energyat r --+ 3p. becomesmore andmore negative(i.e.,lorverandlou,er)asr
[Link] relationshipbetweenoutu'ardand inu,ardforcesallou's
one to concludethat the electronshouldrnovefasteras it movescloserto the
nucleussincer'2 : Ze7l(rntr). [Link]'herein this mclclel is a conceptthat
relatesto the experimental fact that eachatoll erltitsonll'certainkindsof pho-
C h a r a c t e r i z a t i o no f [Link] u'[Link] occurreduficn an [Link] irr
small and large stable a largercircularorbit lostenergyandnrovedto a [Link] [Link].r'ever.
orbits of radii 11 and 12 the Newtoniandynamicsthatproducedthe [Link] allou'orbrtsof
for an electron moving
arouno a nucleus. any [Link] henceany [Link] be [Link]"it wouldappcartlratthe
electronshouldbe ableto emit photonsof any encrgyas it movedfi'omorbit to
orbit.
The breakthroughthat allowed scientistssuch as Niels Bohr to apply the
circular-orbitmodel to the observedspectraldata invol','edfirst introducinsthe
ideathat the electronhas a wavelengthand that this u'avelensthi is relatedto
its nromentumby the de Broglie equationL- hlp. The key stepin the Bohr
model u'asto also specifythat the radiusof the circularorbit be suchthat the
cilcurnference of the ctrcle2nr equalan integer(n) multipleof theu'avelen-gth i.
Only in this way will the electron'svu'ave
experienceconstructiveinterferenceas
the electronorbits the [Link], the Bohr relationshipthat is analogousto
the Bragg equationthat determinesat what anglesconstructiveinterferencecan
occuris

2 t r r: n ) . . (1.7)

Both this equationandthe analogousBraggequationare illustrationsof what we

call boundaryconditions;they are extra conditionsplacedon the wavelengthto
producesomedesiredcharacterin the resultantwave( in thesecases,constructile
interference).Of course,there remains the questionof why one must impose
theseextra conditionswhen the Neu,toniandynamicsdo not requirethem. The
resolutionof this paradoxis one of the thingsthat quantummechanicsdoes.
R e t u r n i n g t ot h ea b o v ea n a l y s i as n du s i n gl , : h l p : h l ( m v ) . 2 r r : 4 ) ' , s g
weli as the force-balanceequationm"r2 1, : Z e2I 12, one can then solvefor the
radii that stableBohr orbits obey:

y :1nhl2n)2 /(m"Ze2) (l 8)

and, in turn. for the velocities ofelectrons in these orbits,

v : ze2lfuhl2tr). (l . e )
 W h y q u a n t u mm e c h a n i c si s n e c e s s a r fyo r d e s c r i b i n gm o l e c u l a rp r o p e r t i e s

Thesetrvo resultsthen allorv one to expressthe sum ofthe k i n e t i c( l r r . r ' : ) a n c l

Coulombpotential(-Ze2 lrl energies as

g : - ! r , , z t r ' 1 t n l1t 2 t1 ' . (1.r0)

) '

Just as in the Bragg diflraction result,rvhich specifiedat what anglesspecial

high intensitiesoccurredin the scattering, there are many stableBohr orbits.
each labeledby a value of the integer ir. Those with small n have small
radii. high velocitiesand more negativetotal energies(n.b., the reference
zero of energycorrespondsto the electronat r : oc, and with v :0). So.
"allowed" that causes
it is the resultthat only certainorbits are only certain
energiesto occur and thus only certain energiesto be observedin the emitted
photolls.
It turnedout that the Bohr formula for the energylevels(labeledby r) of
an electronmoving abouta nucleuscould be usedto explainthe discreteline
e m i s s i osnp e c t roaf r l l o n e - e l e c t r o
ant o m sa n d i o n s{ i . e . .H . H e . L i - : . e t c . ) t o
[Link] suchan interpretation of theexperimental [Link]
that a photonof energy

hv=R(tlni-rlni) ([Link])

is emittedrvhenthe atom or ion undergoesa transitionfiom an orbit having

quantumnumberni to a lower-energy orbit havingnf. Herethe symbolR is used
to denotethe fbllo*ing collectionof factors:
I
R - -nt,Z-e*llltr)tl' (1.12)

The Bohr formulafbr energylevelsdid not agreeaswell w ith the observedpattern

of emissionspectrafor speciescontainingmorethana [Link],
it doesgive a reasonable fit, for [Link] the Na atomspectraif oneexamrnes
only transitionsinvolvingthe [Link] primary reasonfor
the breakdorvn of the Bohr formulais the neglectof electron-electron Coulomb
repulsionsin its derivation. [Link] success of this modelmadeit clear
thatdiscreteemissionspectracouldonly be explainedby introducingtheconcept
thatnot all orbitswere"allowed".Only specialorbitsthatobeyeda constructive-
intert-erence conditionwerereallyaccessible to the electron'[Link] idea
that not all energiesrverealloweclbut only certain "quantized" energiescould
occur was essentialto achievingevena qualitativesenseof [Link]
experimentalfact that emissionspectrawerediscrete.
In summary,two experimentalobservationson the behaviorof electronsthat
were crucial to the abandonmentof Newtoniandynamicswere the observations
of electrondiffractionand of [Link] of thesefindings
seem to suggestthat electronshave some wave characteristicsand that these
waveshaveonly certainallowed(i.e.,quantized)wavelengths.
 T h e b a s i c so f q u a n t u mm e c h a n i c s

So. nou, rve havesome idea why the Newton equationsfail to accountfor
the [Link] light and small;lartrclessuchas electronsand nuclei.
We seethat extraconditions(e.g..the Braggconditionor constraints on the de
Brogiieu'avelength) couldbe imposedto achievcsome degree of agreenrenlu'ith
experimental observation. u'estill arelcft [Link]
Hou'e'u'er,
are that can be appliedto properlydescribesuch [Link] and why the extra
conditions areneeded. It turnsoutthata ner.r'kind of equationbasedon cornbining
u'aveandparticleproperties neededto be developed to address suchissues. These
arethe so-calledSchrddinger equationsto ri'hichwe no\\'tunr our attention.
As I [Link] one hasvet shownthat the Schrodi-qer equationfollou,s
deductivelirfrom [Link] is. scientistsdid not de-
theypostulatedit. Sorre ideaof hou'the scientists
rive this equatiotr; o1'thatera
"dreamedup" the Schriidinger equationcan be had by [Link] tirre and
thc
spatialdepcndence thatcharacterizes so-called travelingu,aves. lt shouldbe noted
that the peoplervhoworkedon theseproblcmskneu'a greatdcal aboutu'aves
(e.g.,soundw'avesand \\.ater\\'aves)and thc equationsthey [Link].
they kneu' that u,avescould sometimesdisplaythe characteristic of quantized
u'avelenqths or fi'equencies (e.9..fundamentals and overtonesin soundu'a'"'es).
They'kneu',for [Link] r.r,a'u'esin one dinrensionthat are constrainedat trvo
points(e.g..a violin stringheld fixed at tu'o ends)undergooscillatorymotjon
in spaceandtime with characteristic fi'equencies andu'avelengths. For [Link].
the motion of the violin stringjust nrentionedcan be describedas havingan
arnplitudeA(r. t ) at a position-r alon-qits len-sthat tinle I gircn by

l ( , 1 -/ .) : , 4 ( . r - . 0 ) c o s ( 2 . ' r y r ) . (1.13)

where r, is its [Link] arnplitude'sspatialdependence

alsohas
a sinusoidaldependence givenby

I (,r.0) : A sin(2r:;l).1. (1.1.1)

where.),is the crest-to-crest

lengthof the [Link] examplesof suchwaves1n
onedimensionareshownin Fig. 1.4. In thesecases,the stringis fixedat x : 0 and
at x : l, so the $,avelengthsbelongingto the two wavesshownare ), : 2L and
),= [Link] the violin string were not clarnpedat ir: I, the wavescouid have
any value of ),. [Link] string is attachedat ir : l. the allowed
u'avelengthsare quantizedto obey

)"= Lln. (1.15)

sin(hx/LJ

w h e r en : 1 . 2 , 3 . 4 , . . . T h e equation
t h a t s u c hw a v e so b e y ,c a l l e dt h e w a v e
F u n d a m e n t aaln d f i r s t [Link]
overtone notes of a
d2,4(x,t) . d2A
v i o l i ns t r i n g . (r . l 6 )
Lll- (l\-
 W h y q u a n t u mm e c h a n i c si s n e c e s s a r fyo r d e s c r i b i n gm o l e c u l a rp r o p e r t i e s

wherec is the speedat which the [Link] speeddependson the com-

positionof the materialfrom which the violin string is [Link] the earlier
expressions for the -r- and t-dependences
of the wave,A(x,t). we find that the
rvaves frequencyandwavelengtharerelatedby the so-calleddispersionequation:

v : : ( c l n ) :. (r .r 7 )
or

c: tv. (1.18)

This relationshipimplies,for [Link] an instrumentstring made of a very

stiff material(largec) will producea higherfrequencytonefor a givenwavelength
(i.e.,a givenvalueof n) thanwill a stringmadeof a softermaterial(smallerc).
For waves moving on the surface of, for example. a rectaneulartwo-
dimensionalsurfaceof lengthsI,, and 1,,, one finds

l(.r. r. /) = sinln,:rr I L.,)sin(n,:ry

I L, )cos(2zur) (1.19)

[Link] wavesare quantizedin two dimensionsbecausetheir wavelengths

mustbe constrained to cause,4(x.y. r) to vanishat_r : 0 and-r : Z. aswell as
at -r,- 0 and,y : I,, for all times t. Let us now return to the issueof wavesthat
describeelectronsmoving.
The pioneersof quantummechanicsexaminedfunctionalforms similar to
[Link],fbrmssuchas I : exp[*22 i( ur - .y/i)] were
consideredbecausetheycorrespondto periodicwavesthat evolvein -vanclt under
no external,t'-or [Link] that

Jt.J /22 \r
;(/.\-
- : - l . l l (1.20)
\ ,\ ,/
andusingthe de Brogliehypothesis
X: hlp in the [Link] finds

,l:.1 .71.2 \l
-:-p-l , | ,a (r.21)
.1.\'- \ tt /

If I is supposedto relateto the motionof a particleof momentump underno

externaltbrces(sincethe waveformcorrespondsto this case),p2 canbe related
to the energy E of the particle by E : p212m. So, the equationfor I can be
rewrlttenas

'ti1= -2,,r('* ,,.

(1.\'- n
)' \ /
(1.22)

or, aiternatively.

-(*)'#: uo r t 1l\
10 T h e b a s i c so f q u a n t u mm e c h a n i c s

Returningto the time-dependence

of l(.r'. /) and using v : Elh. one can also
show that

/ h \ dA
i ( ^ I ' : E i t ' (1.24)
\tlr / dl

which, using the first [Link]

th 1d.4 r lt \:J:A
'\=/ :- (1.2s)
,t, \;/ ,..

This is a primitiveform of the Schrcidinger

equationthatwe rvill addressin much
more detail [Link], what is importantto keep in mind is that the use of
the de BroglieandPlanclCEinstein connectionse,: h/p an<iE : lr,), bothof
which involvethe constanti. producessuggestive connectionsbetrveen

,(+)# and
E (r . 2 6 )

and bet."veen

r , 1/t1:d:
D- anc ( 1 . 2| 7
\ 2:r) dt:
or, [Link]

7.' and - , (:-'\ +

\ :1T / .1.\
(1.18)

Theseconnectionsbetweenphl,sicalproperties(energyf and mornentump)and

differentialoperatorsare someofthe unusualfeaturesofquantum mechanics.
The abovediscussionabout wavesand quantizedu,avelengths as rvell as the
obser"'ationsabout the wave equationand differentialoperatorsare not meant
to provide or even suggesra derivationof the Schrodin_eer [Link] the
scientistswho inventedquantummechanicsdid not deriveits rvorkingequations.
Insteadthe equationsand rules ofquantum mechanicshavebeenpostulatedand
designedto be consistentu'ith [Link] studentsoften find
this to be disconcertingbecausethey arehopingand searchingfor an underlying
fundamentalbasisfrom vvhichthe basiclarvsof quantummechanicsfollou, log-
icall1,.I try to remind them that this is not hou' theory u,[Link] uses
experimentalobservationto postulatea rule or equationor theory,.and one then
teststhe theory by making predictionsthat can be testedby further experiments.
Ifthe theory fails, it must be "refined", and this processcontinuesuntil one has
a betterand [Link] this [Link],with all of its un-
usual mathematicalconstructsand rules. shouldbe viewed as arising from the
imaginationsof scientistswho tried to invent a theory that was consistentwith
experimentaldataand which could be usedto predict thingsthat could then be
testedin the [Link] far, this theory hasprovenreliable,but, of course,
 T h e S c h r o d i n g eer q u a t i o na n d i t s c o m p o n e n t s t l

"new and improved"

we are ahvayssearchingfor a theory that describeshow
smalllightparticlesmove.
If it helpsyou to be more acceptingof quantumtheory,I shouldpoint out
that the quantLrm descriptionof particleswill reduceto the classicalNewton
description undercertaincircumstances. In particular,when treatingheavypar-
ticles(e.g.,macroscopic massesand evenheavieratoms),it is oftenpossibleto
useNer,[Link]. rve will discussin more detail how the quantum
and classicaldynamicssometimescoincide(in which caseone is free to use
the simpler Neu,tondvnamics).So. let us now move on to look at this [Link]
Schrodinger equationthat lve havebeendi-eressingaboutfor so long.

1.2 The Schrodingerequationand its components

It hasbeenwell established that electronsmoving in atomsand moleculesdo
not obeytheclassicalNewtonequations of [Link] agotriedto treat
electronicnlotionclassicallv. and tbundthat featuresobservedclearlyin erperi-
mentalmcasurements simplywerenotconsistent with sucha treatment. Attentpts
r,veremadeto supplcrnent the classicaleqr-rartions rvith conditionsthat could be
usedto rationalizesuchobservations. For [Link]
theangularmomentllmL : r x p be allowedto assumeclnl,vintegermultiplesof
l/22 (lvhichis ofien abbrcviated as fi), which can be shorvnto be cquivalenr to
the Bohrpostulatetti : 2:rr. Hor,vever. until scientistsrealizecl
thata ner',,
setof
[Link] cluantum [Link] light rnicroscopic particles.a rvide
[Link] laboratoryobserlationsof molecule-lcvel phenomena and
thc equations uscdto describesuchbehavior.
Quanturnrnechanics is castin a languasethat is not farniliarto moststudents
of chemistryrvhoare exarrininsthe subjectfbr the first [Link] nrathernatical
contentandhorvit relatcsto experimental measurements bothrecluirea -grc-at
deal
of eftortto [Link] thesethoughtsin mind.I haveorganizedthis ntaterialin
a mannerthatfirst providesa brief introductionto the two prirnaryconstructs of
qLlantum mechanics operators andwavefunctionsthatobeya Schrodinger equa-
[Link].I dcmonstratc theapplication of theseconstructs to severalchemically
[Link] learningthe solutionsof the Schrodinger eqLlatlon
fbr a f'ewmodel systems,the str"rdent can betterappreciatethe treatmentof the
fundamental postulates of quantummechanics as well astheirrelationto experi-
mentalrreasurement for whichthewavefunctionsof theknolvnmodeloroblems
off!r inrportuntinrerpr,.'rations.

1. 2 . 1 O o e r a t o r s
E a c hp h y s i c a l l ym e a s u r a b l eq u a n t i t yh a s a c o r r e s p o n d i n go p e r a t o rT, h e
e i g e n v a l u eosf t h e o p e r a t o tre l lt h e o n l y v a l u e so f t h e c o r r e s p o n d i n pg h y s -
ical propertythat can be observed.
12 T h e b a s i c so f q u a n t u mm e c h a n i c s

Any experimentallymeasurablephysicalquantityF (e.g.,energy.

dipole mo_
[Link] angularmomentum,
linearmomentum,
kineticenergy)hasa classicalmechanicar expression in termsof thecartesianpo_
sitions{q;} andmomenta{p; } of theparticlesthatcomprisethe
systemof interest.
Eachsuchclassicalexpression is assigned
a correspondins quantummechanical
operatorF formed by replacingthe lp1) in the classicalform
by the differentiar
operaror-i/1 0/0q [Link] reavingthe coordinates97 that appearin F untouched.
[Link] classicar kineticenergyof y'y'
particles(u,ithmasses r?1)movlnir
in a potentialfield containingboth quadraticand rinear
coordinate-def"nd.n.e
can be written as

r:,I, + t, ( q,- q) ,) l
t/2k( q1- ,tl) (1.29)
l t'i /2 u ,,+
The quanturnmechanicaloperatorassociatedwith this
.F is

t :,I, - +t'(q,-'ti')]
L*,# * )r (,' q;')' (I . 3 0 )

Suchan operatorwould occurwhen,for example,one

describes the sum of'the
kineticenergies of a collectionof particles(the
f,:, ,,(7ti/2,,11rerm).plus rhe
sum of "Hooke'sLavu'''parabolic potentialstthe lfZy,,=t.t k(qr _ q!t'721, ana
(the last term in F) the interactionsof the particles
*itn ln .*[Link] appried
field whosepotentiarenergyvarieslinearryasthe particles
moveaivayfrom their
equilibriumpositions{4,0}.
Let us try more exampres. The sum of the:-components of angularmomenta
(recallthat vectorangurarmomenfurnL is defined
as L : r x p; ota collection
of ly' particleshas the following classicalexpression:

t: GiPvt-)'tP,i), (r.31.)
f
-/= I ..'\'

and the correspondingoperatoris

t : -t,.,:.Iu
(',* - r',+) ( r. 3 2 )

If one transformsthesecartesiancoordinatesand
deri'ativesinto polar coordi_
nates,the aboveexpressionreducesto

a
F--iltf ( 1. 3 3 )
i?' 3Qi
The r--componentof the dipole moment for
a coilection of 1y'particleshas a
classicalform of

'p -- ./--
\- Z1exi, ( r.34)
 T h e S c h r o d i n g eer q u a t i o na n d i t s c o m p o n e n t s 12

for which the quantttmoperatoris

F: I Z,exi, (r .3s)

*'here Z 1e is the chargeon [Link] that in this case,classicaland

quantumforms are identicalbecause-F containsno momentumoperators.
The mappingfrom F to F is straightforwardonly in termsof Cartesiancoor-
[Link] map a classicalfunction .E,given in termsof curvilinearcoordinates
(evenifthev are orthogonal),into its quantumoperatoris not at all straightfor-
ward. The mappingcan alwaysbe done in terrnsof Cartesiancoordinatesafter
which a transtbrmationof the resultingcoordinatesand differentialoperatorsto
a curvilinearsystemcan be performed.
The relationship of thesequantummechanical operators
to experimental
mea-
surementlies in the eigenvaluesof the [Link] such operator
h a sa c o r r e s p o n d i negi g e n v a l ueeq u a t i o n

F x ,: u , x , (r . 3 6 )

in which the X7 arecalledeigenfunctions andthe (scalarnumbers)cv;arecalled

eigenvalues. All sucheigenvalue equations areposedin termsof a givenoperator
(F in this case)and thosefunctions{X/} that F actson to producethe function
back againbLrtmultipliedbv a constant(the eigenvalue). Becausethe operator
F usuallircontainsdifferentialoperators (coming from the momentum).these
equations aredifferentialequations. TheirsolutionsX, dependon thecoordinates
thatF containsasdiflerentialoperators. An examplewill helpclarif,vthesepoints.
The differcntialoperatord f d.r' actson rvhatfunctions(of_r,)to generatethe same
functionbackagainbut multipliedby a constant? The answeris functionsof the
form exp(rl-r,)
since

(1.37)

So. we saythat exp(a_r,) is an ei-eenfunctionof d ldy and a is the corresponding

eigenvalue.
As I will discussin moredetailshortly,theeigenvalues of theoperatorF tell us
the onlt,valuesofthe physicalpropertycorrespondingto the operatorF that can
be observedin a laboratorymeasurement. SomeF operatorsthat we encounter
possess eigenvalues that arediscreteor [Link]
measurement will result in only those [Link] F operatorshave
eigenvalues that can take on a continuousrangeof values;for theseproperties,
laboratorymeasurement can give any value in this continuousrange.
14 T h e b a s i c so f q u a n t u mm e c h a n i c s

1 . 2 . 2W a v ef u n c t i o n s
The eigenfunctiono s f a q u a n t u m m e c h a n i c aol p e r a t o rd e p e n d o n t h e
coordinatesupon which the operator acts. The particularoperator that
c o r r e s p o n d st o t h e t o t a l e n e r g yo f t h e s y s t e mi s c a l l e dt h e H a m i l t o n i a n
[Link] eigenfunctionsof this particularoperator are called wave
functions.

A specialcaseofan operatorcorresponding to a phvsicalll,nreasurable quan-

titv is the HamiltonianoperatorH that relatesto the total enercyof the system.
The energyeigenstates ofthe svstemW are functionsofthe coordinates1g;1
t h a t H d e p e n d so n a n d o f t i m e l . T h e f u n c t i o ni V ( r 7 ; . 1 ) l l : W - V g i v e st h e
probabilitydensitvfor observingthe coordinates at the valuesq; at tirnel. For
a manv-particle systemsuchas the H2o molecule"the u'avefunctiondepends
o n m a n yc o o r d i n a t e F
s .o r H 2 O .i t d e p e n d o
s n t h e . x . . r .a
. n d : ( o r r . 0 . a n dS 1
coordinates of thetenelectronsandthe.r-..r,.and: (or r. 0. and(t) coordinates oi'
theoxygennucleusandofthe two protons;a totalof39 coordinates appearin v.
ln classicalmechanics. the coordinatcs q7 and their correspondins momenta
p.i arefunctionsof [Link] stateof the systemis thendescribed bl,spccifvine
q j(l) andp 1([Link] quantumntechanics.theconceptthat17;is knor,"'n asa funr:tion
of tirneis replacedby the conceptof the probabilitvdcnsitvfor 1indingQi ar ?
particular"'alueat a particulartime lv(r7r. t)lr. Knorvledge of thc corresponding
momentaas functionsof time is alsorelinquished in quantumnrechanics: lgain.
onlyknou'ledge of theprobabilitvdensiryfor finding7r, [Link],,aluc
at a particulartime / remarns.
The Hamiltonianeigcnstates are especiallyirnportantin chenristrybecause
manyof thetoolsthatchemistsuseto studymoleculesprobetheenergystatesof
[Link], nrostspectroscopic methodsaredcsigned to deterrnine
u,hichenergystatea moleculeis in. [Link] melh-
odsthatmeasureotherproperties(e.g.,the :-cornponentof angularmonrentun.r
or the total angularmomentum).
As statedearlier,if the stateof somemolecularsvstemis characterized b1'a
wavefunctionw that happensto be an eigenfunction of a quanlummechanical
operatorF, one can immediatelysay somethingabout what the outcome $,ill
be if the physicalproperty F correspondingto the operatorF is [Link]
particular,[Link]

Fxt : )",x1. (r . 3 8 )
wherei 7 is oneof theeigenvalues of F. we know thatthevaluei7 wili be observed
if thepropertyF is measuredwhile thernoleculeis describedby thervavefuncrion
'4t:
[Link] fact,oncea measurement of a physicalquantityF hasbeencarried
out and a particulareigenvalue).i hasbeenobserved-the system'srvavefuncrion
v becomesthe eigenfunction1, that correspondsto that eigenvalue. That is, the
 T h e S c h r o d i n g eer q u a t i o na n d i t s c o m p o n e n t s 15

actof makingthe measurement causes the system'swavefunctionto becomethe

eigenfunction of the propertythat tvasmeasured
What happensif someotherpropertyG, [Link] quantummechanicaloperator
is G is measuredin sucha case?We knou' from what was said earlierthar some
ei_genvaltrepr of the operatorGlvill be observedin the measurement. But. will
the molecr-rleiwavefunctionrenrain,after G is measuredthe eigenfunction of
F, or will the measurement of G causeV to be alteredin a way that makesthe
molecule'sstateno longeran eigenfunction of F? It turns out that if the two
operatorsF and G obey the condition

F G = G F (1.39)

[Link] the propertyG is measuredthe wavefunction'Q : Xtrvili remain

unchanged. This [Link] two operators
does
not [Link] calledcontmutation:that is, we savthe tu,ooperatorscommuteif
[Link] us seehowthispropertyleadsto theconclusionabout
V remainingunchangedif the tr,[Link] particulaqwe apply
theG operatorto the aboveeigenvalue equation:

C F 7 . ,: G 1 i Xi . ( r. .10)
[Link] comntutation to re-writcthe leti-handsideof this [Link]
u s et h e f a c tt h a ti , i s a s c a i a n
r u r n b e r t ot h u so b t a l n

F G7i: r.Q7, (1.-1lt

So. [Link] scc that (G;1,) itself is an ei-uenfunctionof F havingeigenvalue

ir. So, unlessthereare [Link] one eigenfirnctions of F corresponding to
i; (i.c..unlessthis cigenvalue
the [Link] is degenerate).GX, [Link]
proportionalto [Link] write this proportionalirv conclusionas

GX,: ltix,. (1.:12)

which meansthat X, is also an eigenfunctionof G. This. in turn. meansthat

measuringthe propertyG rvhilethe systemis describedby the wave function
tp :
X i d o e sn o t c h a n g et h ew a v et u n c t i o ni:t r e m a i n sX , .
So. lvhenthe operatorscorresponding to two physicalpropertiescommLlre,
once one measures one of the properties(and tlius causesthe systemto be an
eigenfunction ofthat operator).subsequent measurcment ofthe secondoperator
will (iithe eigenvalue of the first operatoris not degenerate) producea uniqtLe
eigenvalue ofthe secondoperatorandwill not changethe systemwavefunction.
If the two operatorsdo not [Link] simply can not reachthe above
conclusions. In [Link] of the propertycorresponding to the
first operatorrvill leadto one ofthe eigenvalues ofthat operatorand causethe
system lvave function to become the [Link],
to T h e b a s i c so f q u a n t u mm e c h a n i c s

subsequent measurementof the secondoperatoru'ill producean eigenvalueof

that [Link] the systemu'avefunction will be changedto beconrean crgen-
functionofthe secondoperatorand thusno longerthe eigenfunction ofthe first.

1 . 2 . 3T h e S c h r o d i n g eerq u a t i o n
T h i s e q u a t i o ni s a n e i g e n v a l u ee q u a t i o nf o r t h e e n e r g yo r H a m i l t o n i a n
operator;its eigenvaluesprovide the only allowed energy levels of the
sysrem.

The ti me-depe n dent eq u ati o n

l f t h e H a m i l t o n i a no p e r a t o rc o n t a i n st h e t i m e v a r i a b l ee x p l i c i t l yo, n e m u s t
solvethe time-dependentSchrodingerequation.

Beforemovingdeeperinto understanding whatquantummechanics "means".

it is useful to learn horv the wave functions V are found by applying the ba-
sic equationof quantun.r mechanics. the Schrodinger [Link] a few exactly
[Link] solutionsto these"easy"vet chernically
very relevantmodelsr.l,illthen facilitatelearninsmore of the detailsaboutthe
structureof quantummechanics.
The Schrodinger equationis a differentialequationdepending on tirneandon
all ofthe spatialcoordinates necessarl,to describerhesystemat hand(thirt1,-nine
for the H2O examplecited above).lt is usuallyn,ritten

H t U : i 7 16 r y 1 ' 6 , . (1.43)

wherev(qi, t.) is the unknownwave function and H is the operatorcorrespond-

ing to the total energyof the [Link] operatoris called [Link]
and is formed, as statedabove,by first writing down the classicalmechanical
expressionfor the total energy(kinetic plus potential)in Cartesiancoordinates
and momenta and then replacingall classicalmomentap; by their quantum
mechanicaloperatorsp j : -ih 0lSq.i.
FortheH2o exampleusedabove,theclassicalmechanicalenergyof all thirteen
particlesis

F _ [Link]; l/r\-e -\-2,"'t

+ l r ^ - 't - /_J t.t.t ? ,," I

- r+- fl pr j* -r-' -r 'ir- \-? Z'Zr'i

ruh
I ( r.44)
I'
wherethe indicesi andj are usedto label the l0 electronswhose30 Cartesian
coordinatesarelq il anda and b labelthe threenucleiu,hosechargesaredenoted
{Zo}, and whosenine cartesiancoordinatesare lqo}. The electronand nuclear
 T h e S c h r o d i n g eer q u a t i o na n d i t s c o m p o n e n t s
17

massesare denotedrr. and {n,,}, [Link] correspondingHamiltonian

operatoris

H=E
* #l . : l * - + T l
.+[-*#.:+'#l .'15)
NoticethatH is a secondorderdifferentialoperatorin thespaceofthe 39 Cartesian
coordinatesthat describethe positionsofrhe ren electronsand [Link] is
a secondorder operatorbecausethe momentaappearin the kinetic energyas pr2
andp;,and the quantummechanicaloperatorfor eachmomentump : _ f ttl 0q
I
is of first [Link] the H2o example at hand then
reads:

't7 l-Ai- - 1\- "' - \- Z,e:w

1 2n,dq; ??r,, ?,.,"
-tl_ l a: ;l r!
'll2n,'dq.; * 1 l Z , Z n e :\ y : ln- ( 1.16)
2? rr.h At

The Hamiltonianin this casecontains/ [Link] exampleof a casewhere11

doescontain/ occurswhenan oscillatingelectricfield E cos(ror)alongthe-r-axrs
interactswith the electronsand nuclei and a term

)- f
Z,rX,,[Link]((r/ e;r7E
cos(rr-rr) (1.47)
)
is addedto the [Link],xo and.,videnotethe.r coordinatesof the
ath
nucleusand the7th [Link].

The ti me-i ndepe ndent eq uatio n

lf the Hamiltonianoperator does not containthe time variableexplicitly,

one can solve the time-independentSchrodingerequation.

In caseswherethe classicalenergy,and hencethe quantumHamiltonian,

do
not contarntermsthatareexplicitlyrimedependent (e.g..interactions with time
varying externalelectric or magneticfields rvould add to the above
classical
energyexpressiontime dependentterms),the separations of variablestechniques
can be usedto reducethe Schrodingerequationto a time-independent
equation.
In suchcases,H is not explicitlytime dependent, so one can assumethat
v (.qi ' t ) is of the form (n.b.,this stepis an exampleof the useof the
separations
of variablesmethodto solvea differentialequation)

V ( q , , t ) = W ( qi ) F ( t ) (l.4E)
18 T h e b a s i c so f q u a n t u mm e c h a n i c s

this "ansatz"into the [Link]-dependent

Substituting equationgives
Schrodin_ser

V ( q 1 )i t t 0 F l ' d t: F Q ) H V ( qt t . (1.49)

D i v i d i n gb y V ( q 1 1 F ( tt)h e ng i v e s

F t t i ht ) F l i t r t : W i [ H , ! ( 4)i] . (1.50)

SinceF(l ) is only a functionof time t. andtU(q 1)is only a functionof the spatial
coordinates lq 1], andbecausethe left-handand riglit-handsidesmust be equal
for all valuesof t and of lq il, both the left- and right-handsidesmust equala
constant. lf this constantis calledE. the two equationsthatareernbodiedin this
separatedSchrodingerequationreadas follou,s:

HV(q,): EV(q,). ( l . 5 l)
ifidF(t)ldt: EF(t). [Link])
Thefirstoftheseequations is calledthetin-re-independent Sclrrodinger equatron: it
is a so-called eigenr,alue equationin u'hichoneis askedto find functionsthat1,ield
a constantmultipleof [Link] when actedon by the Hamillonianoperator.
Suchfunctionsarecalledeigenfunctions ofH andthecorresponding coustallts are
c a l l e dc i g . e n r a l u oe fsH . F o re x a m p l ei .f H u e r eo f ' r h ef o r m ( _ l i ' 2 . \ lt ; t :, d Q : -
H" tlrenfunctionsof thc form exp(inrQ)would be eigenfunctions because

I li d: I ln:h:l
,-
- - l
ztra6,1e x p ( i', , i @ ) : l * f crptiirrd,). ll.-i.l)
I [:,r,J
In [Link] 2tt2
12tvtis theeigenvalue. In [Link]
the squareof an angularmomentumoperator(recall earlierthat u,'eshou'edtlre
:-cornponent of angularmomentumis to equal-ih d ldQ).
\Vhen the Schrcidingerequation can be separatedto -seneratea time-
independent equationdescribingthe spatialcoordinatedependence ofthe u'ave
function,the ei_eenvalue E mustbe retumedto the equationdeterminingF(t) to
find the time-dependent part of the [Link] solving

ifldF(r)ldt: EF(t) (1.54)

once E is [Link]

F(t) = gapl-i E1171, (1.55)

and the full wave function can be written as

v ( qi . r ) : w ( q i )e x p ( -i E t l t t ) . (1.56)

For the aboveexample,the time dependence

is erpressedby

F(r) : exp(-irlnt2ti1zll1itt1. (l . s 7 )
 T h e S c h r o d i n g eer q u a t i o na n d i t s c o m p o n e n r s
tv

In summary,wheneverthe Hamiltoniandoes not dependon time explicitly,

one can solvethe time-independent Schrodinger equationfirst and thenobtain
rhetime dependence as exp(-i El/fi) oncethe energyE is known. In the case
of rnolecularstructuretheory,it is a quite dauntingtask even to approximately
soivethe full Schrodinger equationbecauseit is a partialdifferentialequarion
dependingon all of the coordinates of the electronsand nucleiin the molecule.
For this [Link] are various approximationsthat one usually implements
when attemptingto study [Link].

The Born-O ppen hei me r approxi m ati on

one of the mostimportantapproximationsrelatingto applyingquantummechan-
ics to moleculesis knownas the Born-oppenheimer(Bo) approximation. The
basicideabehindthis [Link] the full electrons-
plus-nucleiHamiltonianoperatorintroducedabove,

H : t l - , iz n tri t'q--: l \ - - " ' - ' t a l

?1 2?'',, ?,,"1
7 7 . :
z!/ zxc
*,-l_fir;rr_l \..-
(r . 5 8 )
" 2nr,,;)q,] 2 ';
t'u.h
|
the tirne scalesivith which the electronsand nuclei move are generallyquite
difl'[Link] [Link] heavynuclei(i.e.,evena H nucleusweighsnearly
2000timeswhatan electronweighs)move(i.e.,vibrateandrotate)moreslowly
thando the lirrhterelectrons. Thus.r','cexpectthe electronsto be ableto ..adjust,'
theirmotionsto themuchrrroreslowlymo'n,ing [Link]
[Link] solvethe Schrcidingerequationtbr the movementof the electronsin the
presenceof fixed nuclei as a vvilvro representthe fully adjustedsrateof the
electronsat any fixedpositionsofthe nuclei.
The electronicHamiltonianthatpertainsto the motionsof theelectronsin the
presence of so-calledclarnpednuclei.

H : -\ - l - l 'r tit , ": t -, 1*: 1 t ' ' ' - r'Z{

t/ l (1.59)
I )L,"., ,,, l'
producesas its eigenvalues.
throughthe equation

H O 1 ( q1 i r l , ) : E r k l , , ) , ! . t ( q I, q , ) . (1.60)
energresErGl,) that dependon where the nuclei are located(i.e., the
{r7,,}
coordinates).As its eigenfunctions. one obtainswhat arecalledelectronicwave
functionsltLrkti I r7,,))which alsodependon wherethe [Link]
energiesE xQ") arewhatwe usuallycall potentialenergysurfaces. An example
of sucha surlaceis shoivnin Fig. [Link] surfacedependson two geometrical
coordinates{r7,,}a'd is a plot of one particulareigenvalue
EL@,) vs. thesetwo
coordinates.
20 T h e b a s i c so f q u a n t u mm e c h a n i c s

Secondorder saddlePotnl
fransrtion structureA

Minimum for
product A

- 0.,5

Two-dimensional Secondorder n
saddlepoint
potentialenergysurface
\ralley-ridgc
s h o w i n gl o c a lm i n i m a , 0.5
i n f l e c t i o np o t n t
transitionstatesand Minimum for reactant
p a t h sc o n n e c t i n g
them.

Although this plot has more information on it than we shall discussnou'.

a few featuresare worth [Link] be tlireeminima (i.e..points
of -87 rvith respectto both coordinatesvanishand u'here
r,l,herethe derir"atives
the surfacehas positivecurvature).Thesepoints correspond"as we u'ill seeto-
u,ard the end of this introductorymaterial,to geometriesof [Link]
(labeledtran-
[Link] surfacealso displaystwo first order saddlepoints
sition structuresA and B) that connectthe three mininra. Thesepoints havezero
first derivativeof E.r with respectto both coordinatesbut have one direction
of [Link] we will show later, these points describetransition
statesthat play crucial roles in the kineticsof transitionsamongthe three stable
geometries.
Keep in mind that Fig. 1.5showsjust one of the tJ surfaces;eachmolecule
has a ground-statesurface(i.e., the one that is lowest in energ.v)as well as an
[Link]'snou'returnto our discussionofthe
BO model and ask what one doesonceone has suchan energysurfacein hand.
The motions of the nuclei are subsequently,within the^Bo model. as-
sumed to obey a Schrodingerequationin u'hich L"{-(h'l2nt,,)d!l\q;-l
1l2l,n ZoZ6e2f ro.6]* E r@) definesa rotation-vibrationHarniltonianfor the
particularenergystateE5 [Link] rotationalandI'ibrationalenergiesand
wave functionsbelongingto eachelectronicstate(i.e., for eachvalue of the in-
dex K in E x@)) arethen found by solving a Schrodingerequationrvith sucha
Hamiltonian.
This BO model forms the basis of much of how chemistsview tnolecular
For exarnple,as appliedto forrraldehyde
structureand rnolecularspectroscopy.
H2C:O, we speakof the singletgroundelectronicstate(with all electronsspin
paired and occupyingthe lowest energyorbitals) and its vibrational statesas
 M o t i o n o f a p a r t i c l ei n o n e d i m e n s i o n 21

nell as the n --+ z* and ir --->lr*electronicstatesand their vibrationallevels.

Althoughmuch more will be said about theseconceptslater in this text. the
studentshould be arvareof the conceptsof electronicenergysurfaces(i.e.. the
lExe,)]) and the vibration-rotarion statesthatbelongro eachsuchsurface.
Having been introducedto the conceptsof operators,wave [Link]
Hamiltonianandits schrodin-eer equation,it is importantto now considerseveral
examplesof the applicationsof [Link] examplestreatedbelorvwere
chosento providethereaderwith vaiuableexperience in solvingtheSchrodinger
equation:they werealsochosenbecausethey form the n]ostelementarychemical
modelsof electronicmotionsin conjugated moleculesand in atoms,rotationsof
linearmolecules,and vibrationsof chemicalbonds.

1.3 Yourfirst applicationof quantum mechanics- motion

of a particlein one dimension
This is a very important problem whose solutionschemistsuse to model
a wide variety of phenomena.

Let's begin by examiningthe motion of a singleparticleof mass,n ln one

directionwhich rvewill call "r rvhiler-rnder the influenceof a potentialdenoterl
Z(,r). The classicalexpressionfor the total energyof such a systemis .E :
p212n -t V(x), rvherep is the momentumof the particlealong ther-axis. To
focuson specificexamples, considerhorvthis particlewould moveif I/(.r.1il'ere
of the forms [Link] Fig. I .6. wherethe total energyE is denotedby theposition
ofthe horizontalline.

1 . 3 . 1C l a s s i c aplr o b a b i l i tdy e n s i t y
I would like you to imaginewhattheprobabilitydensitywould be for this particle
moving with total energyE and with I/(,r) varying as the threeplots in Fig. L6

Three
characteristic potentials
showing left and right
c l a s s i c at lu r n i n gp o i n t s
at energiesdenotedby
t h e h o r i z o n t aIli n e s .
22 T h e b a s i c so f q u a n t u mm e c h a n i c s

[Link] conceptualize theprobabilitydensitl'.imaginethe particleto havea

[Link] to it andthink of this lampblinkingsay100timesftrreach
time intervalit takesfor theparticleto completea full transitfrom its lefi turning
point to its right turning point and back to the [Link] turning points,r-sand
-rRarethe positionsat whichtheparticle,if it weremo'"'ing underNe\\'ton'siau's.
wouldreversedirection(asthe momentumchangessign)andturn [Link]
positionscan be found by askingwherethe momentumgoesto zero:

0 : p : ( 2 m ( E - I ' ( . t ) )r r. (1.6i)

Thesearethe positionsrvhereall ofthe energyappearsas potentialenergvE :

I/(.x) and correspondin the abovefiguresto the pointsrvherethe dark horlzontal
linestouchthe I'(r)plots as sho'uvn in the centralplot.
The probabilitydensityat any valueof -r:represents the fractionof'tirllethe
particlespendsat this valueof (i.e..
..r- u'ithin.r and,t * r/.r).Think of forrrling
this density'by allowingthe blinking lantpattachcdto the particleto shedlight
on a photo-eraphic plate that is exposedto tliis light for many oscillationso1-
the particlebet\\'een -r1 and;rp. [Link] can expressthis probability
amplitude P(r) by dividingthe spatialdistancer/.r"by thevelocitl'of theparticle
at tlie nointr:
I r r r rr i r
P ( - t ): ( 2 m ( E - I " ( . r) ) (i.62)

BecauseE is constantthroughoutthe particle'smotion.P(,t)tl'ill be smallat,t

valuesu,heretheparticleis mo'u'ingquickly (i.e..'ui'here
l/ is lou') andri'ill be high
wherethe particlemovesslowly 1r,r,here tr/is high). So. the photographic plate
u'ill shou,a brightregionwhere I/ is high (because the particlemovesslou'lyin
suchregions)and lessbrightnesswhere tr' is lou'.
by analyzingthe
The bottomline is thatthe probabilitydensitiesanticipated
classicalNewtoniandynamicsof this oneparticle'uvould appearasthehistogran.r
plotsshownin Fig. 1.7 [Link](and
thuslower Z(.lr)),it spendslesstime and P(.t) is [Link]

Classical
ty plotsfor the
three potentialsshown
i n F i g .1 . 6 .
 M o t i o no f a p a r t i c l ei n o n e d i m e n s i o n 23

[Link] spendsmore time and P(.r) is [Link] the plot on the ri_eht,/(,r)
"box".
is constantwithin the so the speedis [Link](.r) is constant
fLrrall .v valuesrvithinthis one-dimensionalbox. I ask that you keepthesepiots
in nrind becausethey are very difrerentfrom what one finds when one solves
the Schrcidinger equationfor this [Link] pleasekeepin mind that
these plots representwhat one expectsif the particle were moving accordingto
classicalNewtoniandynamics(w,hichwe know it is notl).

1 . 3 . 2O u a n t u mt r e a t m e n t
To solve tbr the quantLlmmechanicalwavefunctionsand energiesof this same
problem,we first write the Hamiltonianoperatoras discussedaboveby replacing
n Ov -lll(l /Ll.\".

h' ,t:
Il:-'--l+l'rrr (1.63)
lilt tl.\-

We then try to find solutionsry'(.r)to Hrlr : ErL that obey certainconditions.

Theseconditionsare relatedro the fact that iry'(.r)12 is supposed to be the prob-
ability densityfbr findinethe particlebetrveen -r and.r * r/.[Link] keepthingsas
simpleas possible,let'sfocuson the "box" potential/ shorvnin the right side
o f F i g . 1 . 7 .T h i sp o t e n t i a le. x p r e s s eadsa f u n c t i o no f , r i s : V ( x \ : c c f o r - r < 0
andfbr,r > L: I/1.r1: 0 fbr.r-between0 and l.
The factthat tr/is infinitefor-r < 0 andfor.r > l. andthatthe totalenergyE
mustbe [Link] /r mustvanishin thesetwo regionsdt : 0 for_r < 0 and
forx > l). This conditionmeansthat the particlecan not accessthesereslons
wherethe potentialis [Link] [Link] makeuseof is that
/(-t ) mustbe continuous: thisrrreans thatrheprobabilityof theparticlebeingrt,r
cannot be discontinuously relatedto theprobabilityofit beingat a nearbypornr.

1 . 3 . 3E n e r g i e sa n d w a v e f u n c t i o n s
The secondorderdift-erential
equation

il tlllr
- -;-. -t tr/\r)tlr : Etlt (1.6.+)
^
lnt dr'

hastwo solutions([Link]
it is a secondorderequation)in the regionbetrveen
"r:0andx:L:

/ : s i n ( l . r ) a n d r y :' c o s ( , t - r ) . r v h e r e A i s d e f i n e dka :s( 2 m E l t t : 1 | i 1 . ( 1 . 6 5 )

[Link] mostgeneralsolutionis somecombinationof thesetrvo:

ty': .1sintA.r)
+ B cosrfr.r'.y. (1.66)
z4 T h e b a s i c so f q u a n t u mm e c h a n i c s

T h e f a c t t h a t r y ' m u s t v a n i s h a t , v : 0 t n . b . . r / r v a n i s h e s<f o0r xa n d i s c o n t i n u o u s .

so it mustr.'anish at tlrepoint.r :0)means that the rveithtinganrplitudeof the
cos(Ax)termmustr,'anish becausecos(Ar): I at-t : 0. That is.

B =0. (1.67)

The amplitudeof the sin(rt,r)rerm is not affecredby the conditionthat {r vanish

ot.r : 0. sincesin(t.r) [Link]
at,r : 0. [Link],weknou,that{r is really
of the forn:

/(.r):.{ sin(tr-). (1.6[i)

The conditionthat 1/ralsovanishat -v : I hastrvopossibleimplications. Either

. 4 : 0 o r f r m u s tb e s u c ht h a t s i n ( A l ) : 0 . T h e o p t i o nI : 0 w o u l dl e a dt o
an answerry'that vanishesat all valuesof -r and thusa probabilitythat vanishes
everywhere. This is unacceptable because it rvouidimplv thattheparticleis never
observedanywhere.
The otherpossibilityis that sin(ft1): 0. Let'sexplorethis ansrver becausert
offers the first exanrpleof energyquantizationthat you have probablyencoun-
[Link] you [Link] sinefunctionvanishes at integralmultiplesof :r. Hence
AL mustbe somemultipleof :r: let'scall the integern andu'riteZA : rz (using
the definitionof fr) in the form:

Lt}nEllitt): nn. (r.69)

Solving this equation for the energy f , we obtain:

E : n1r2rt
l(2mt). (r . 7 0 )

This result saysthat the only energyvaluesthat are capableof grving a wave
function ry'(x)that will obey the aboveconditionsare thesespecificE [Link]
otherwords,not all energyvaiuesare"allowed"in the sensethattheycanproduce
ry'functionsthat are continuousand vanishin regionswhere z(x) is infinite. If
one usesan energyE that is not one ofthe allowedvaluesand substitutesthis E
into sin(*;r),the resultantfunctionwill not r,anishat x : L.I hopethe solutionto
this problemremindsyou of the violin stringthatwe [Link]
the violin stringbeing tied down at x : 0 and at -r : z gaverise to quantization
of the wavelengthjust asthe conditionsthat ry'be continuousat;r-: 0 [Link] : Z
gaveenergyquantization.
Substitutingk : nn /L into ry': I sin(k,r)gives

VQ)= Asin(nrx/L) (r.7r)

The valueof I can be found by rememberingthat lq/ l2 is supposedto represent
theprobabilitydensityfor finding the particleat x. Suchprobabilitydensitiesare
 M o t i o n o f a p a r t i c l ei n o n e d i m e n s i o n 25

supposedto be [Link] their integralover all x valuesshould

amountto unity. So, we can find I by requiringthat
f r

t: l l t t r ) l i , / . t: l . -1t : s i n 2 t r xz I L ) t t x . 0.72)
J ;l
where the integral rangesfrom r : 0 to x : L. Looking up the integral of
sin:(a,rI and solving the aboveequationfor the so-callednormalizationconsranr
I gtves

A : (2lDtt2 (1.73)

andso

/(-r) : (2111trt sin(nrx I L) (r . 7 4 )

Thevaluesthatn cantakeon aren : l, 2. 3. . . .; the choiceie : 0 is unaccentable
becauseit would producea wavefunction ry'(.t)that vanishesat all r.
The full ,r- and t-dependentwavefunctionsare then given as

v ( , r . r ) : ( : ) ' r i n ' ' l ' r' . r l - . t ' : : ' ! l (r.75)

\ L / L lttl L-/lt
L )

Noticethatthe spatialprobabilitydensitylv1,r. r)12is not depenclent

on timeand
is equalto lry'(.r)12
becausethe complexexponentialdisappears when v*w is
[Link] meansrhatthe probabilityof findingthe particleat variousvalues
of "r is time-independent.
Anotherthing I wantvou to noticeis that,unlikethe classicaldynamicscase.
notall energyvaluesE [Link]. In theNewtoniandynamicssituation,E could
be specifiedandthe particie'smomentumat any-r valuewas thendetermined to
within a [Link] [Link] quantummechanics onemustdetermine,by solving
the Schrcidinger [Link] allowedl'aluesof 6 are. TheseE valuesare
quantizedmeaningthattheyoccuronly for discretevalues.t : ,2 r2h)
/(2r, L2)
determinedbv a quantumnumberrr. by the massof the particlem, andbychar-
acteristics ofthe [Link] this case).

1 . 3 . 4P r o b a b i l i t yd e n s i t i e s
Let's now look at sorne of the wave functionsv(.r) and compare the proba-
bility densitieslv(.r)ir that they representto the classicalprobabilitydensities
[Link] n : 1 and n : 2 wave functions are shown in the top
of Fig. [Link] corresponding probabilit-v
densitiesare shownbelowthe wave
functionsin two tbrmats (as n--vplots and shadedplots that could relateto the
flashinglight way of monitoringtheparticle's locationthatu'e discussed earlier).
A more completesetof [Link](for izrangingfrom 1 to 7) are shownin
F i e .1 . 9 .
zo T h e b a s i c so f o u a n t u mm e c h a n i c s

,ffi'ffi
n:2
The two
lowest wave functions
and probability
densities.

{
V.

m,axffita;,;*r

25 mau,*ffiu,,*;r
Node

t6

9
Seven
lowestwave functions I

andenergies.
I
 M o t i o n o f a p a r t i c l ei n o n e d i m e n s i o n
27

Noticethat as the quantumnumbern increases, the energyE also increases

lquadratically rvith n in this case)andthe numberof nodesin v also increases.
Also notice that the probability densitiesare very different from what we en-
countered earlierfor [Link][ [Link] at ther : I andil : 2
densities and comparethem to the classicaldensityillustratedin Fig. [Link]
classical densityis easyto understand becausewe arefamiliarwith classicaldy-
[Link] this case,we saythat P(.r) is constantwithin the box because the fact
that tr/(.r) is constantcauses rhekineticenergyandhencethe speedofthe particle
to [Link] [Link] r : I quantumu'avelirnction'sp(.r) plot is
peakedin the rniddleof the box and falls to zeroat the [Link] n :2 density
hastwo peaksP(-r) (oneto the leti of the box [Link] one to the right).a
C l a s s i cIa
nodeat the box [Link] falls to zeroat the [Link] thing that students p r o b a b i l i t yd e n s i t yf o r
"how
oftenaskme is doesthe particlegerfrom beingin the lefi peakto beingin p o t e n t i a l s h o w n .
theright peakif it haszerochanceof everbein_q at the midpointwherethenode
is?"The difficr"rlty rviththis questionis rhatit is posedin a terminologytharasks
for a classicaldynamicsanswer. [Link] asking"horvdoestheparticleget. . ."
one is dernandingan answerthat involvesdescribingits motion (i.e..it moves
from hereat time t1 to thereat time 11).Unfortunatell,. quantummechanics does
notdealwith issuessuchasa particle's trajectory(i.e..rvhereit is at varioustinres)
butonly vu'ithits probabiltyof beingsomewhere (i.e.,lw l:;. The nc'xtsectionwill
treatsuchparadoxicalissuesevenfirrther.

1 . 3 . 5C l a s s i c aaln d q u a n t u mp r o b a b i l i t d
yensities
[Link],it is temptingfor mostbeginningstudentsof quanturnmechanics
to attemptto lnterpretthe qr-rantum behaviorof a particlein classicalterms.
[Link] adventureis tull of daneerandboundto fail becar-rse [Link]
particlessimply do not moveaccordingto Nervton'[Link] [Link] try
to "undersrand"r'"'hatkind of (classical)motion wor-rldbe consistenrlvith the
ri : I or n :2 quanturnP(_r)plotsshownin Fig. [Link] I hope
1,ou
[Link] gainingclassical understanding of a quantumresultrvill
not "work" in that it r'"'illleadto nonsensical [Link] point in leadingyou ro
attemptsucha classicalunderstanding is to teachyou thatclassicalandquantum
resultsaresimplydifferentandthatyou mustresisttheurgeto imposea classical
understanding on quantumresults.
For the zr: I case,we note that P(x) is highestat the box midpoint and
vanishesat.r : 0 and,r : L. Lna classicalmechanicsworlcl this would mean
that the particle movesslowly near-y : L/2 andmore quickly near-r : 0 and
r : L. Becausethe particle'stotai energyE must remain constantas it moves.
in regions where it moves slowlv. the potential it experiencesmust be high.
and where it moves quickly, tr/ must be small. This analysis (n.b., basedon
classicalconcepts)would leadus to concludethat the n : I p(i arisesfrom tne
 28 T h e b a s i c so f q u a n t u mm e c h a n i c s

particlemoving in a potentialthat is high near.r : L

/2 anrlloll, as-r:approaches
0orZ.
A similaranalysisof the n :2 P(x) plot would leadus to concludethatthe
particlefor which this is the correctp(_r) rnustexperiencea potentialthat
is
high midwaybetween:r: 0 and x : L/2,high midu,ayberu,een .r : L 12 and
x = L , a n d v e r y l o w n e a r x: L / 2 a n d n e a r - r: 0 a n d r : r . T h e s e c o n c l u s i o n s
are"crazy"becausen'e know that the potentialt,(-r)for which we solvedthe
Schrodin-eer equationro generatebothof thewavefunoions (andbothprobability
densities)is constantbetween-r : 0 andt : L. That is. ."veknow the samel,(-r
)
appliesto theparticlemovingin thell : I andr : 2 states. whereastheclassical
motion analysisofferedabovesuggeststhat l,(l ) is differentfor thesetwo cases.
what is wron-qrvith our attemptto understandthe quantump(.r) prots?
The
mistakewe madewas in attemptingto apprythe equatio'sandconcepts
of clas-
sicaldynarnicsto a P(.i-)protthat did not arisefrom classicarmotion.
Siniply
put' one can not ask how the particleis mo'ing (i.e..what is its
speedat vari-
ouspositions)whentheparticleis undergoing [Link] students.
r,r'hen
first experiencingquantumrvavefunctionsandquantumprobabilities,
trv to
thinkof theparticlemovingin a classicalway thatis consistenl with thequantum
P(-r). This attemptto retaina degreeof classicalunderstanding of theparticle's
movementis alwaysrnetwith frustration,as I illustratedu,ith the above
examole
andwill illustratelaterin othercases.
continuingwith thisfirstexarnpleof hou'onesolvestheSchrodinger equatlon
and how one thinks of the quantizedE 'alues and u,avefunctionsv,
let me
offer a little more optimisticnote than offeredin the precedingdiscussion.
If we
examinethe w(.-v) plot shownin Fig. 1.9fern :7, andthi'k orthe corresponding
P(;r) : lv(x)12,we notethatthe p(.r-)plot rvouldlook somethinglike that
shown
in Fig. [Link] would havesevenmaximaseparated by six [Link] we wereto prot

I
0.9 11 iI h
0.8 I \ 1 I
0.7 r
^. (\A
5
r! 0.5

V o.c I
I
I
0.3
0.2 I I
0.1 I
I
Ouantum
probabilitydensityfor I I
n : 7 s h o w i n gs e v e n
peaksand six nodes.
0 v rJ \
xlL
 M o t i o n o f a p a r t i c l ei n o n e d i m e n s i o n 29

: 55,we would find a P(,r) plot having

1V (.r) lr for a very larger valuesuchasn
55 maxima separatedby 54 nodes,with the maxima separatedapproximately
" c o a r s eg r a i n e d "
b ) , d i s t a n c eosf ( 1 / 5 5 I ) . S u c ha p l o t ,w h e nv i e w e di n a sense
1i.e.. focusinu with somewhat blurred vision on the positions and heights of
the maxima) looks very much like the classical P(,r)plot in which P(.r) is
constantfor all -r. In fact, it is a general resultof quantummechanicsthat the
quanfumPt.r) distributionsfbr largequantumnumberstake on the form of the
classicalP(-r) for the samepotentialZ that was usedto solvethe Schrodinger
[Link] is also true that classicaland quantum results agree when one is
dealingwith heavyparticles. Forexample,a givenparticle-in-a-box energvCn :
n:h: I \2ntL2) u,ouldbe achievedfor a heavierparticleat highern-valuesthanfor
a [Link],heavierparticles,movingwith a givenenergyE. have
higherr and thusmoreclassicalprobabilitydistributions.
\!'e rvill encounter thisso-calledquantum-classical correspondence principle
'nve
again "r,'hen examineother model [Link] is an important property of
solutionsto the Schrodinger equationbecauseit is what allowsus to bridgethe
"gap" betrveen usingthe Schrodinger equationto trearsmall,light particlesand
theNewtonequationsfor macroscopic (big, heavy)systems.
Anotherthing I u'ould like you to be arvareof concerningthe solutionsry'
and .E to this Schrodinger equationis that eachpair of wavc functionsry',,and
r/r,,belongingto difl-erent qltantumnumbersn andn'(and to differentener_qiesl
displaya propert,vterrnedorthonormality. This propertymeansthat not only are
tfr,,and,{r,, eachnormalized
r t
l: ltrlt,,:,/t:/ .a,,:lt'. (1.16)
J , J

but they are also orthogonalto eachother

o= lr',t,,,r',t,,,
a, (1.77)

where the complex conjugate* of the first function appearsonly when the ry'
solutionscontainimaginarycomponents (you haveonly seenone suchcasethus
far - the exp(ill @) eigentunctions
of the:-componentof an-qular
momentum).It
is commonto write the integralsdisplayingthe normalizationand orthogonality
conditionsin the fbllorvineso-calledDiracnotation:

| - / r l r ,|, 1 1 , , \ 0: 1 r l t , ,r' ! , , ) . (1.78)

where the l) and (l si'mbolsrepresenttb and lt*, respectively,and putting the

two togetherin the (l) constructimpliesthe integrationoverthevariablethatry'
dependsupon.
The orthogonality conditioncanbe viewedas similarto the conditionof two
vectorsv1 and v2 beingperpendicular. in which casetheir scalar(sometirnes
called"dot") productvanishes .
Vl vr : 0. I wantyou to keepthispropertyin mind
5U T h e b a s i c so f q u a n t u mm e c h a n i c s

because not only of theseparticle-in-

you r.l'illsoonseethat it is a characteristic
a-box wave functionsbut of all w,ar.,efunctionsobtainedfrom any Schrodinger
equatlon.
In fact, the orthogonalityproperty is evenbroaderthan the abovediscusston
[Link] turnsout that all quantummechanicaloperatorsforrnedas discussed
earlier(replacingCartesianmomentap by the corresponding -ihd ldq operator
and leat'ingall Cartesiancoordinates as thev are) can be to be so-called
shor'vn
Hermitianoperators. This rneansthat they form Hermitianmatricesu'henthey
are placed betweenpairs of functionsand the coordinatesare integratedover.
For example,the rnatrix representation of an operatorF vu'henacting on a set of
functionsdenoted{dr} is

F rL : ( Q t t F t Q: L t * ; r r ' u , (1 . 7 e )
|
For all of the operatorsformed following the rules [Link] finds that
thesematriceshavethe followingproperty:

rr J _I -r +r / t . r1.80)

rvhichmakesthe matricesu'hat we call [Link] tlre functionsupon u'hich

F actsand F itseifhaveno imaginaryparts(i.e..arereal).thenthe tnatricesturn
out to be svmmetric:

F r . L: Ft.r ( r . 8 )l

The importanceof the Hermiticity oI symlnetryof thesematriceslies in the fact

[Link] canbe shownthat suchmatriceshaveall real(i.e.,not [Link])
eigenvalues
and haveeigenvectors that are orthogonal.
So, all quantummechanicaloperators,not just the Hamiltonian,have real
eigenvalues(this is good since theseeigenvaluesare rvhat can be measuredin
any experimentalobservationof that property) and ortho-qonaleigenfunctions.
It is importantto keep thesefacts in mind becausewe make use of them many
times throughoutthis text.

1 . 3 . 6T i m e p r o p a g a t i o no f w a v e f u n c t i o n s
For a systemthat existsin an eigenstateV(x) : (2lL)t12 sn(nitx lL) havingan
energyEn : n2r2h2lQmLz),the time-dependent u'avefunctionis

r): (?)"',inff*r(-?),
v(x. (1.82)

which can be generated by applying the so-called time evolution operator U(l' 0)
to the wave function at I : 0:

V(x.r) : U(r,0)v(,r,0) (1 . 8 3 )
 M o t i o n o f a p a r t i c l ei n o n e d i m e n s i o n

rvherean explicitform for U(t, t') is

L ( r .r ' \ : . * o[ - " ' *'''"-l. (1.8+.)

L n l
Thc-functionV(.r, t) has a spatialprobabilitydensitythat doesnot dependon
timebecause
v.(.r. : (;) ""' (?)
r)v(.r.,, (1 . 8 5 )

sinceexp(-rtE,'l/i) exp(itE,,l/i) : l. However it is possibleto preparesystems

(evenin real laboratorysettings)in statesthat are not singleeigenstates;we call
suchstatessuperposition [Link] example,considera particlemovingalong
within the "box" potentialbut in a statewhosewave function at some
the-r--axis
initialtime/:0is
, / l \ l : / 1 7 r \ . . r 2 t l : / l - z . r ' \
u(ror:,,,(;)'''( {r86)
. )-,,(;),'"(?)
T h i s i s a s r - r p e r p o s i t i o n o f t h eIna:n d z : 2 e i g e n s t a t e s . T h e p r o b a b i l i t v d e n s i t y
a s s o c i a t endi t h t h i s f u n c t i o ni s

r,r,r' | (?) - r(?)

: I ;,'",(?) +/ sin:
lrr\
' t ' n /( / 1 . - r .tr I
(1.87)
r /t'"( . ,/i
Then : I andn : 2 components, the superpositionV. andthe probabilityden-
:
sityat 1 0l V lr areshownin thefirstthreepanelsof Fig. l. 12. It shouldbe noted
that the probabilitydensityassociared with this superposition srateis not sym-
metricaboutthe"r : I/2 rnidpointeven thoughthen : I andn : 2 component
wavefunctionsand densities [Link] densitydescribes the particlelocalized
more stronglyin the large-"vregionof the box than in the small-,rregion.
Now. let's considerthe superpositionwave function and its densitl,at later
times. Applying the time evolutionoperator exp(-itHl/i) to V(r,0) generates
thistime-evolved tunctionat time /:
",'"
' { i ) ' , , " ( T ) - ' ' '(?r)
v ( ;rr1 *: ' ( - # )
{, (?)l
:
['
' ( ; )" ' ' ( ? ) ' - io) (- -' ' ,'G)"'
"',"(?).-'(-+)] (r.88)

The spatialprobabilitydensityassociatedwith this W is

,)r::
rv(r j{(;),'",(+)
. (;),,'',(?)
-'(?).o, - u,)if,',
(+),'"(?)] o8e)
[ru,
5Z
T h e b a s i c so f q u a n t u mm e c h a n i c s

I u.l
d o.E r I--r\ 0 ^ --
/ i I \
E t,.o
/f 06 !_t
\ f_' t I
l i . . - lt .l 0.i
I
.- 04
(1.-l / l
_0.1 I
^
{ { ()l t
> {)l
t
t
= = 0 I
\ j
( 1I.
-t ! 7
-0:
X,L

1.6
{ ^l i. { -1^
t: t.l / \ I
f \
J 1 6l
I
I
0 N i \
.\
0.6
t:
.=
I
I
1).1 I a
: d a
|I,
vi )
L_] -

jflff r h e n : 1 a n d n : 2 w a v e f u n c t i o n st,h e i r
s u p e r p o s i t i o na,n d t h e
a n dt i m e - e v o l v ep
d r o b a b i l i t yd e n s i t i e so f t h e
superposition.

At r : 0' this functionclearlyreducesto that

u,rittenearlierfor v(r.0). Notice
thatastimeevorves. thisdensitychanges because of thecos[(82- E 1ltf til factor
it contains'In particular,note that as r moves
through a period of rength6l :
n h/ ( E 2 - f r ). the cosfactorcha'gessign.
Thatis. for r : 0. the cosfactoris + l
f o r l : r / i / ( E 2 - t r ) , t h ec o sf a c t o r i s_ l ; f o r : 2 r / i l ( E z _ ;
t E r ) . i t r e t u r n st o
* L The resultof this time variationin the cosfactor
is that lV l: .lrun-c.,in forn.r
from that shownin the bottom left paner
of Fig. l .12 to tr-ratshou,n in the bottom
right panel (at t : trril(E2 - 6r)) and then
back to the fbr'i in the bottom reft
panel(at t : 2tr11/(Ez- E
t)). one caninterpretthistime variationasdescribing
the particlekprobabilitydensity(not its
classicar positionl),initiaily Iocalized
toward the right side of the box, moving
to the left and then back to the rr_qht.
of course,this time evolutionwill continue
overmore and more cycresas trnre
evolves.
This exampleiilustratesonce againthe
difficulty with attemptingto locarize
particlestliat arebeingdescribed
by quantumwave-functions. For exampre.a par-
ticle that is characterized by the eigenstate (2/ D) /2 sin(tr x I L) i, rn"rill*.,u ,"
be detectedfleor'i : L/2thannearir :0
orrj - r becausethe squareof this
f u n c t i o n i s l a r g e n e a r=- rL / 2 . A p a r t i c l e i n t h e s t a t e
(2lDt/2rin1:n..7tj,smort
l i k e l i z t ob e f o u n d r & r x: L / 4
a n dx : 3 L f 4 , b u t n o t n e a r , r: 0 , x : 1 1 t . s y
 M o t i o n o f a p a r t i c l ei n o n e d i m e n s i o n J5

.r = L. The issueof horv the particle in the latter statemovesfrom being near
r = L f 4 to x : 3 L l4 is not [Link]
mechanicsdoesnot allow us to follow the particle'[Link] is what we
needto know whenwe askhow it movesfrom oneplaceto [Link].
superposition wavefunctionscan offer"to someextent,the opportunityto tbllow
the motionof the [Link] example,the superpositionstatewritten aboveas
2 - r , : 1 2 1 1 1 1s1i n2( l z - r l Z )- 2 - t / 2 Q l L ) t / 2 s i n ( 2 n x l L ) h a s a p r o b a b i l i t y a m p l i -
tude that changeswith time as shown in the figure. Moreover,this amplitude's
major peakdoesmove from side to side within the box as time [Link], in
this [Link] ciln say with what frequencythe major peakmovesback and forth.
"follorv"
In a [Link] [Link] to the particle'[Link] only to the
extent that'"ve are satisfied with ascribingits locationto the positionof the major
peakin its [Link] is, we can not really [Link] "precise"
[Link] we can follor,vthe location of where it is very likely to be found.
This is an importantobservationthat I hopethe studentwill keepfreshin mind. It
is alsoan importantingredientin modernquantumdynamicsin which localized
wavepackets,similarto superposed eigenstates, are usedto detailthe position
andspeedof a particle'smain probabilitydensitypeak.
Thc-aboveexampleillustrateshow one time-evolvesa \ryavefunctiontharcan
be expressed as a linearcombination(i.e..superposition) of eigensrates of the
problem at hand. There is a large amount of current eftbrt in the theoreti-
cal chemistrycommunityaimedat developingelicient approximations to the
exp(-i t H lll) evolutionoperatorthatdo not requireV(.r. 0) to be explicitl_y rvrit-
tenasa sumof eigenstates. This is importantbecause, for mostsystemsof direct
relevanceto molecules,one can not solvefor the eigenstates; it is simply too
difficult to do so. You can find a significantlymore detailedtreatmentof this
subjectat the research-level in my TheoryPageweb siteand my el,tIC tertbook.
[Link]'sspenda little time on a brief introductionto what is involved.
The problemis ro expressexp(-itHlh)V(qr), whereV(qr) is some initial
wave function but not an eigenstate,in a manner that does not require one
to first find the eigenstates{V7} of H and to expand V in terms of these
eigenstates:

v:fc,w, ( 1.e0)
J

afterwhich the desiredftlnction is written as

:
"t (-#) v@,)
1,,*,*,(-+) ([Link])

The basicideais to breakH into its kinetic I andpotentialtr/energycomponents

and to realizethat the differentialoperatorsappearin I only. The importance
of this observationlies in the fact that I and Z do not commute,which means
thaLTV is not equal to ZI (n.b.,for two quantitiesto commutemeansthat their
34 T h e b a s i c so f q u a n t u mm e c h a n i c s

order of appearancedoes not [Link]).Why do they not conrmute'lBecauseI

containssecondderivatives n'ithrespecttothecoordinates {q , } that I' dependson,
[Link] example,d2I d q2LI,@)V (q )l is not equalto Il (q ) d2i d (12V (r7). The fact
that r and l' do not commuteis inrportantbecausethe rnostcommon approach
to approxrmating exp(-rlH/fi) is to rvritethis [Link] in rerms of
exp(-itTlh) and exp(-lri.'/fi). [Link] identity

*'(-f):*'(-+)
*'(-f) (l q2)

is not fully valid as one can seeby expanding all threeofthe aboveexponential
f a c t o r sa s e x p ( . r ) : I * . r - + x 2 / 2 ! * . . . . a n d n o r i n _tqh a l t h e r $ , os i d e so f t h c
aboveequationonly a-eree if one canassulnethatTl': I'L u,[Link] rvenoted-
is not true.
In [Link] approaches to time propagation. onedividesthe time interval
r into nrany(i.e..P of them) small time "slices"r : t / p. onc then expresses
the evolutionoperatoras a productof P short-timepropagators:

itH\ rrHl
e x /p ( - , , : . * o /l - - ) . - o ( -I ; , i) t. H- n\ ( - r i r H t
/ ; )
irtl\lr
: I t
(r.e3)
L ' * n \ -t / l
If one can developan efficient rneansof propagatingfor a short time r. one
can then do so over and over again P times to achievethe desiredfull-time
propagation.
It can be shown that the exponentialoperatorinvoll'ing 11 can betterbe ap-
[Link] in terms of the f and I/ exponentialoperatorsas follows:

rrH\ . -t f- t-'F- V
/ \-
exp l - ln. *\ p ( -/ irr'\ r irr\
(r.e_r)
ft /-.*pt,-. a /.-o(_;/
So. if one can be satisfiedwith propagatingfor very short tirne [Link]
(so that
the rl term can be neglected),one can indeeduse

*'(-#)-*p(-+)-'(-T) (r.e5)

as an approximationfor the propagatorU(r, 0).

To progressfurther, one then expressesexp(-irT//1) acring on the initial
function v(q) in terms of the eigenfunctionsof the kinetic energy operator
r. Note that these eigenfunctionsdo not dependon the nature of the poten-
tial z, so this stepis valid for any and all [Link] ei_eenfunctions
of r :
-n-
*1 ,^ rt , , f

/ tntd'/ctq' arc

, ! , , ( q ) : ( * )*' "r ( T ) ( 1.e6)

 M o t i o n o f a p a r t i c l ei n o n e d i m e n s i o n
35

andthey obeythe following orthogonality

f
V i , t a t t l r , , ( q t :d q3 r p - p l (r . 9 7 )
J
relations
andcompleteness

rL,.t,t \lp = |rq - q t.

trlr,",tt1 (r . e 8 )
J
Writing V(q) as

v ( q ) : I A u- ( ] ' ) v ( q ' ) d q , . (1 . e 9 )

for d(q - q') gives

andusingthe aboveexpression

rr(q) : )rtt(q'ytq'ttp.
v,,u,,,t';,,q' (1.100)
| |
Tlreninsertin-etheexplicitexpressions
for tlr,(c1)andr!
r'z ;(q' ) in termsof ,,ft 1:
,,1c1
0 l2ir 1t pr7
exp(i /fi) -erves
. I tr:
Q t ' 7 t :f f l |
' * o1( ;i l7 r 4( 1; ,7) 1 r :. . - o/ ( -i p# u) t Q r 't1' t ' 1 ' t p r r r 0 r t
JI \; J
[Link]
exp(-i r T lIt)to act on this form for W(r7)produces

. . P/ ( i r T )\ w t q , :r r / ; ' , ' r t i \ /I
\''
n JJ . ' p ( ' # l l r ; t
^ . *' oL[ ? , u : , ,, l f : ) ' ' ' * , r , r o r , r , ( r r 0 2 )
) \ h t h
The integralover p abovecan be carriedout analyticallyanclgives

. * P/(
irT\. 1 m r : / f i n r q _ 4 ' t : \l t t t , t ' ) , t " t
n ) q , , t , : ( : r , r , ) 7 . * oL ," 1
. rr.l03r
So, the final expressiontbr the short-time propagatedwave function is

t f , n t \ ' t f - . ^ f i n t 1 t- 1( t 7 : 1
v"{' ,,r ."r-) ": .u* L
o[-irli'l (.,,r/ {1.10+)
/r-j 7.*pL ,* lVtt1'td,l.
whichis theworkingequationoneusesto computeV (q , t )knowingV (4). Notice
thatall oneneedsto kno."r,
to applythis formulais the potentia1 v (q) ateachpornt
ln [Link] doesnot needto know anyof the eigenfunctions of the Hamiltonian
to apply this [Link]'u'er. one doeshaveto use this formula over and over
agarnto propagatethe initial wavefunction throughmany small time stepsr to
achievefr"rllpropagationfor the desiredtime intervalt : pt.
Becausethistypeoftime propagationtechniqueis a very activeareaofresearch
in [Link] is likely to continueto [Link]
discussionof it is beyondthe scopeof this book, so I will not go further in this
direction.
 T h e b a s i c so f q u a n t u mm e c h a n i c s

1.4 Freeparticlemotions in more dimensions

T h e n u m b e r o f d i m e n s i o n sd e p e n d so n t h e n u m b e ro f p a r t i c r e sa n d
the
number of spatial {and other) dimensions needed to characterize
the
positionand motion of each particle.

1 . 4 . 1T h e S c h r o d i n g eer q u a t i o n
consideran electronof massri and chargee mo'rng on a tlvo-dimensionar
surfacethatdefinesthex, -[Link](e.g.,perhapsan electronis constrained to the
surfaceof a solid by a potentiarthat binds it tightly ro a narro\\,regio' in
the
---directio.).and assumethat the electronexperiences a constantand not trrne-
varvrngpotentialI1t at all pointsin this [Link] pertinenttinre-independent
S c h r o d i n g ec rq u a t i o ni s

-t t /a t a t\..
+ 'drt L i t l t ( x " r ' ) :E f ( ' r " r ' ) (1 10-5)
\a"t' )t(r't')*
The task at hand is to solve the aboveeigen'alueequationto deternrine
"allorved" the
energystatesfor this [Link] termsin this equa_
tion thatcouplemotionin the; and l directions(e.g..no ternlsof the
form,t,-r.b
or Sldx 0/Ev or x 0/8v),separation of'u'ariabres
canbe usedrowrite y'rasaprod-
uct U(r. -r'): l(.r)BLr,). Substitution
of thisforrnintotheSchriidingerequarion,
followedby collectingtogetherall r-dependentandall
terms,
_r'-dependent _qrves
rf 1a]A rt ta2a
2*7ar;- (l'106)
2-E *:E-rtt'
Since the first term contains no r,-dependence and the second contains no
x-dependence. and becausethe right sideofthe equationis independentofboth
x and -r,,both terms on the left must actuallybe constant(thesetwo
consranrs
are denoted E" and Er', respectiveiy).This observationailows two
seDarare
Schrcidinger equationsto be written:

rt . alA
- - / - ' - - F -
(1.107)
ztn dx.

and

- i l b^ - , f B = r ) ' ' (l'lo8)

2m ai
The total energyE can then be expressedin termsofthese separate
energiesE,
and E, as E, + E:, : E - 20. Solutionsto the x- and
.quations
,t,-Schrcidinge.
areeasilyseento be:

A(x):*o[,,.
I r '-l
. | / 2 n E .t ' - |
(+)"'] ano expl-ilt -.1
\ t r /
| (l.t09t
L l
B(,r'):.*p[,r,
(ry)"] a n d . ^ o [ - (, ,+ ) " 1 ( rr o )
L \ r )
 F r e ep a r t i c l em o t i o n si n m o r e d i m e n s i o n s '71

Trio independentsolutionsare obtainedfor eachequationbecausethe,r- and

Schrodinger
.r.-space equationsarebothsecondorderdifferentialequations
(i.e..
a secondorderdifferential
equationhastwo independentsolutions).

1 , 4 . 2B o u n d a r yc o n d i t i o n s
T h e b o u n d a r y c o n d i t i o n s ,n o t t h e S c h r o d i n g e re q u a t i o n ,d e t e r m i n e
w h e t h e rt h e e i g e n v a l u e w
s i l l b e d i s c r e t eo r c o n t i n u o u s .

If theelectronis entirelyunconstrained within the-r. _r'plane,theenergiesE,

and 8.,.can assumeany values:this meansthat the experimenter can "inject"
rhe electrononto the,r.r'plane with any total energyE and any components
8., and ^t.,.alongthe two axesas long as E^ + E;. : E .In sucha [Link]
speaks ofthe energies alongbothcoordinates asbeing"in thecontinuuln"or'.not
quantized".
In contrast. if theelectronis constrained to remainwithin a fixedareain the.r.,r,
plane(e.g.,a rectan-eular or [Link] region),thenthe situationis qualitatively dif'-
i'[Link]-e the electronto any suchspecifiedareagivesriseto boundary
conditionsthatimposeadditionalrequirements on the aboveA anJB fLrnctrons.
Theseconstraints can arise,for example,if the potentiall,ii(.r._r,)becornes verv
Iargetbr,r. .i valuesoulsidetheregion,in r,vhich caserheprobabilityof fincjinethe
electronoutsidetheregionis very [Link] casemightrepresent. for example.
a situationin rvhichthentolecular structurcofthc solid surfacechanges outsidcthe
enclosed regionin a \\'aythatis hi-uhlyrepulsiveto rheelectron(e.g..asin thecase
of molecularcorralson metalstrrfaces). This casecouldthenrepresent a simple
modelof so-called"corrals"in rvhichtheparticleis constrained to a finiteresion
of space.
Forexarnple. if rnotionis constrained to takeplacelvithina rectaneular regron
d e f i n e db y 0 . - r < l , : 0 < , r . < 1 . , . t h e nt h e c o n t i n u i t yp r o p e r t yt h a ta l l w a v e
tunctionsrnustobev (becauseof their interpretation as probabilitydensities.
whichmustbe continuous) causesI ("r) to vanishat 0 andat r , . Thatis. because
.'1mustr.'anish fbr.r < 0 andmustl'anishfor.v > z.r,andbecause A is continuous.
itmustvanishat,r:0andat.r:2.,.Liker.r,ise.B(rr)mustvanishat0andatZ...
To inrplemc'nttheseconstraintsfor I (.r:). onemustlinearlycombinetheabovetwo
s o l u t i o n s e x p [ i , r ( 2Er,rl l i 2 l t i z l a n d e x p [ - r . r ( 2 n E , y / t 2 l r i z l t o a c h i e v e i i f u n c t i o n
thatvanishes at,r : 0:

r(r1:e1p(rf)'"] -.*o "]

([Link] )
l', [-,,{i+)
one is allowedto linearlycombinesolutionsof the Schrodinger
equationthat
havethe sameenergy(i.e.,are de,eenerate)
becauseSchrcidinger
equationsare
Jd
T h e b a s i c so f q u a n t u mm e c h a n i c s

li,ear differentiarequations'An
analogousprocessmust be appriedto g(,r,)
achievea functionthatvanishes to
at.r,: 0.

B(r -op (t+!\"'l -.*" .. - / 1 1 :.1

\ t i I l - q \ fr ) L(12,,r,
' Lf,.', ;/ rnr2
Furtherrequirin-e
J
A(.r) andB(rr) to,,,anish. respecti'elrrat.r : Z.\ and _ 2,.
s r i e se q u a t i o ntsh a rc a nb e o b e y e d _r,
o n l y i f , f , a n d E , a s s u n rpea r t i c u l avr a l u e s :
T
. r n 1 , , , / 2 n Et t : f |
) l - e r n J - i(r; ' ) l = '
(l- ,iarE,1i:l
L r ir . , r
and

f , 2ntE ,l
e r jP ir . ( f i, ' ^ [ l - ' r /' ]\1 1 1 6r , r l
n
L " ) I, J ; L- ) ll : , rrr4r
Theseequationsareequivalent(i.e..
usingexp(i.r): cos-r,* i sin_i,)to

.s, '- n
[ , l r1\2( n- t-EF; -l :)l l- r l a i E ,l r : l
L l = . i n l z(' r i / rr'ir5t
J L J:u
Knou,ing t h a rs i ng v a n i s h e a
ste- nn,forz : l. 2.
3 . . . . l a l t h o u g ht h es i n ( z r )
function'u'anishes for n :0, this function'anishes
for all -t or.r,,andis therefore
unacceptabre becauseit represents zeroprobabiritydensityat alr points
one concludesthat the energiesE, in space).
a'd.E, canassumeo'ly'alues that
obey
t 1,,, tr , l/2
z,{1+rI =r,o.
\ / r / (t.ll6)
. 1 2 n E ,\ r ' 2
r'I ' .l - /l :n,r,
\ f (1.t17)
^r , - nlr2rt
zilr Li (l.rl8)
nln2tl
d
t ff nu d k
t, - = uith
T;8. n\ and t l , = 1 . 2 . 3 , . . . ( l . l1 9 )
It is important to stressthat it is
the imposition of boundary conditions,
pressingthe fact that the electron ex-
is spatia'y constrained that gives
[Link] the absence ,se to
of spatialconfinement"or wjth confinement
only at;r : 0 or Z, or only at y _
0 or 2.,, quantizedenergieswould
realized. rot be
In this exampre,confinementof
the erectronto a finite interval
the 'r and 'r" coordinatesyields energies along both
that are quantizedalong both axes.
the electronis confined"t:lg:l".oJrdinur. If
1.g., u.t*".n 0 < x < 2..) bur not
along the other (i.e., Blu) is eitherrestrictea
ioi,anist aty - 0 or at .1,: L,.
at neirherpoinl).thenthetoralenergy or
E liesin th,
isquantizld
u,i r, i, ""[Link]
orsuch
iinear triatomic morecurehus
;:Tl'.::T;:li.*r;:,TiffT
mo.i than enouj energy in one
of its bonds to
 F r e ep a r t i c l em o t i o n si n m o r e d i m e n s i o n s ?a

ruprureit but not much ener,{yin the other bond; the first bond,sener-evlies in
the [Link]"rtthe secondbond'senergyis quantized.
Perhaps moreinteresting is thecasein whichthebondwith the higherdissoci_
ationenergv is ercitedto a levelthatis not enoughto breakit but that is in excess
ofthe dissociationenergyofthe [Link] this [Link]
states:(i) the stron-sbond havinghi-ehinternalener-qy and the weak bondhaving
low energy(/r). and (ii) the stron_e bondhavinglittle energyandthe weakbond
ha'ing more thanenoughenergyto ruptureit (tL). Althou_eh an experimenrmay
preparethe moleculein a statethat containsonly the former component(1.e.,
rlt : Crltr -f [Link] cr - l. c: : 0), coupringbetrveenthe two degenerate
functions(inducedby terms in the HamiltonianH that havebeen ignoredin
definingry'1and ry'2)usuallycausesthe true wave function W : exp(_il Hih){.,
to acquirea componentof the secondfunctionastime [Link] sucha case,one
speaks of internalvibrationalenergyrelaxation(IVR) givingriseto unimolecular
decomposition of the nrolecule.

1 . 4 . 3E n e r g i e sa n d w a v e f u n c t i o n sf o r b o u n d s t a t e s
For discreteenergy revers,the energiesare specifiedfunctionsthat de-
pend on quantum numbers, one for each degree of freedom
that is
quantized.

Returningto thesituarioni'which motionis constrained [Link]

resultanttotal energiesand wavefunctions(obtainedby insertingthe quantum
ener_cy
levelsinto the expressionsfor '1(,r-)g( r.)) are as follor,",s:

- ,l:iT- tT
' ( r . 12 0 )
2nLi

n:n-lr
t r' -_ :
L

tnt Ll.
- . ( r .r 2 l )

E:E,+8,+tr/o. 1.t22)

y ' (r ,: ( + ) ( # ) ' L * o ( " F- .)* o ( # ) ]

1 | \ : / I \ T , : , ,

.[.-o(
+)-.-,(-i:)]
w i t h n , a n di l , = ! . 2 , 3 , . . . (L.l:l)

The two (1/2Dt/2 factorsare incrudedto guaranteethat ry'is normalized:

w o . 1 , ) l 2 dt txv : I ( L 12,{)
t
NormalizationallowsltL(x , y)r2to be properlyidentifiedasa probabiliry
density
for findingthe electronat a point-lr.r,.
40 T h e b a s i c so f q u a n t u m m e c h a n i c s

rffi)
wl

Plots of the
( a )( 1 , 1 )(,b )( 2 , 1 )(,c )
{ 1 , 2a) n d \ d l Q , 2 lw a v e
f u n c t i on s .

Shou'nin Fig. l.l3 are plots of four suchtwo-dimensional u,avefunctions

f o r r r . a n d n , v a l u e so f ( 1 , 1 ) ,( 2 , 1 ) ,( 1 . 2 )a n d ( 2 . 2 ) .r e s p e c t i ' c l .Nv o
. r e t h a tt h e
functionsvanishon the boundariesof the box, and noticehou,tlre nurnberof
nodes(i.e..zeroesencountered as the r.r,ave functionoscillatesfi-ompositiveto
ne-qati\:e) is relatedto the r- and n.,.quantun nurnbersand to the [Link]
patternof more nodessignifying higher ener_rly is one that u,e encounteragain
and againin quantummechanicsand is somethingthe studentshouldbe able
to useto "guess"the relativeenergiesof [Link] when their plots are at
[Link],you shouldalso [Link] in the one-dimensional box case.
an)'attempt to classicallyinterprettheprobabilities p(-r. corresponding to the
-r,)
abovequanturn\4'avefunctionswill resultin [Link] in the one-dinrensionai
[Link](.:r'. 1,)wouldbe constant alongslicesof fixed.r-andvaryinu_i,
orslicesoffired,r'andvarying-r r.r,ithintheboxbecausethespeedisconstantther
Hou'ever,the [Link] P(x, t,) plots,at leastfor small quantumnumbers,arenot
constant. For larger, andn ,.values,thequantump(r,
l,) plotswill again,via the
quanturn-classical correspondence [Link] (constant)classical
P(.r.,i,)form.
 F r e ep a r t i c l em o t i o n si n m o r e d i m e n s i o n s
41

1 . 4 . 4Q u a n t i z e d
a c t i o nc a n a l s o b e u s e dt o d e r i v e
e n e r g yl e v e l s
Thereis anotherapproachthat canbe usedto find energylevelsand is especially
straightforwardto use for systemswhose Schrodingerequationsare separable.
Theso-called classicalaction(denoteds) of a particremovingwith momentump
alonga pathleadingfrom initial coordinateq1at initial time ri to a final coordinate
q r . a rt i m er ; i s d e f i n e db y

s= (r.125)
/*"'R.o{.
[Link] momentufirvectorp containsthe momentaalong all coordinatesof the
[Link] the coordinatevector q likewisecontainsthe coordinatesalong all
suchdegreesof [Link] example,in thetwo-dimensional particle-in-a-box
problemconsidered above,q : ("r..ir)hastwo components
as doesp : (p,.p,,),
and the action integralis

(r.r26)
In computin-u suchactions,it is essential
to keepin mindthesignof themomentum
as the particlemovesfrom its initial to its final [Link] examplewill help
clarify'thesematters.
Forsystemssuchastheaboveparticle-in-a-box examplefor whichtheHamil-
tonianis separable,the actionintegraldecomposes into a sum ofsuch integrals.
one fbr eachdegreeof [Link] this trvo-dimensional example,the additivity
of H.

H : t i ,- H ,: + - + + v t y t 1v l v t
2n 2n
ti ;;t: - r fi a2
tttt
;:
tt.Y-
I'{.t)- :-:-; * l'( r'). (r.127)
llll dl'-

meansthat p. and [Link] independently solvedfor in termsof the potentials

v(r) and I,'(,u)as well as the energies6. and 6,, associatedwith eachseparate
degreeof freedom:

p, : +r/2i(.1rlo)), (r . r 2 8 )
, / =t t-t l b - , - Y l . v l l ;
Pt= (1.129)

the signs on p, and pr, must be chosento properly reflectthe motion that the
particleis [Link] into the action
rntegralyields

s:s,+,s, (1.13())
fi:rf
f [Link]
: t f2n1E,- v(t))rr * t vtrt))t/.v.(r.l-ll)
J,,,,, Jr,,,, Jrm@,-
42 T h e b a s i c so f q u a n t u mm e c h a n i c s

The relationshipbetweentheseclassicalactionintegralsand the existenceol

quantizedenergylevelshasbeenshownto involveequatingthe classicalaction
for rnotionon a closedpathto an integralmultipleof planck'scorlstant:

s . r n .=
" d f q t = c ' p: [Link] q: r i l r ( , : l . ] . 3 . : 1. . . ) . (1.132)

Appliedto eachof the independentcoordinates

of the trvo-dirnensional
particle-
in-a-boxproblem,this expression
reads
r -'r
l,=L
n . , l t= | , 2 t n ( E , - I ' ( , r ) )d t + | - !f/x(l' Jr(.rD./r. (l ljj)
J.-o - l, = r.
f,=t r f,t
n,h: I t t 2 m ( E -, 1 " ( . r . t) l)r + | * r , 1 i 1 E _ t 1 r y 1, l r . il.13.1)
Jr:0 "lt:t,

Notice that the signs of the niomentaare positivein each of the first inre-
g r a l sa p p e a r i n ag b o v e( b e c a u s teh e p a r t i c l ei s m o ' i n g f i . o m . r : ( ) t o . r - : 1 , .
a n d a n a l o s o u s lfyo r . t - l t t o t i o na. n d t h u sh a sp o s i r i r en r o n t c n t u l lrrn d n e l a r i r t
i n e a c ho f t h e s e c o n di n r e g r a l s( b e c a u steh e r n o t i o ni s f r o m . r : l , t o - i - : 0
(andanalogously for.r'-motion)and thusthe particlehasnegativenromentum).
W i t h i nt h e r e g i o nb o u n d e db y 0 . - r -< l . : 0 < , r . < 2 , . t h c p o t e n t l avl a n i s h e s "
so i/(x) : l'lr') : 0. Using this fact. and re'ersingthe upperanti lou,erlinrits.
andthusthe [Link] the [Link] obtains

n,h:2 [ ' = t ' , ' 7 , , , qd r : 2 r 5 u , P 1 , . (l.li5)

'='
n, h : 2 1Zn t, tlt. : 2{r,, tr, L, (1.3
16)
[

Solving for E, and 8,. one finds

- (n,h12
(r.r37)
6ri l-;

- (n'h\2
(1.138)
s nL l
Theseare the sarnequantizedenergylevelsthat arosewhen the u,avefunctron
boundaryconditionswerematchedatx : 0. r : 2., and,l : 0..1,: 1,,. In this
case!one saysthat the Bohr*sommerfeldquantizationcondition.

nh : [Link], (1.r3e)
lo't=o""
hasbeenusedto obtainthe result.
Theuseof actionquantizationasillustratedabovehasbecomea very important
tool. It has allowed scientiststo make great progresstoward bridging the gap
betweenclassicalandquantumdescriptionsof [Link] particular,
 F r e ep a r t i c l em o t i o n si n m o r e d i m e n s i o n s 43

by usingclassicalconceptssuchastrajectoriesandthenappending quantalaction
[Link] have been able to developso-called
semi-classicalmodelsof
moleculardynarnics. In suchmodels,oneis ableto retaina greatdealof classical
understandingwhiie building in quantumeffects such as energy quantization.
[Link] interf'erences.

1 . 4 . 5O u a n t i z e da c t i o nd o e s n o t a l w a y sw o r k
[Link] approachof quantizingthe action does not alwaysyield the
correcterpressionfor the [Link] [Link] appliedto
the so-calledharmonicoscillatorproblemthat we will study in quantumform
[Link] servesas the simplestreasonablemodelfbr vibrationof a diatomic
moleculeAB. one expresses the totalenergyas

e:{+\,' ([Link])
tlt t

w h e r e 4 : m ^ n t B l 0 nA f r r s ) i s t h e r e d u c e d m a s s o f tAhBe d i a t o m . firs t h ef o r c e
constant describing thebondbetweenA andB. r is thebond-length displacement.
andp is themomentumalongthebondlen-eth. The quantizedactionrequirement
thenreads

, h: I pd :, I l ' , ( ' -1 , ' ) ] ' ' , , (l.l.1l)

i sl c a r r i e do u tb e t w e e ny = - ( 2 8 l k ) l ' l a n dC E l k l ' 1 . t h el e f ta n d
T h i si n t e - u r a
rightturningpointsof [Link] form a closedpath.
Carryingout this integraland equatingitto nh givesthe followin_q expression
for thc energyE:
, , . 1 2
I I I K \
6-:,,. l-l (1.142)
::r \lL/

rvherethe qLlantumnunrberri is allorvedto assumeintegervaluesrangingfrom

n : 0. l. 2. to [Link] problemrviththisresultis thatit is wrong! As experi-
mentaldataclearlyshor'",. thelowestenergylevelsfor thevibrationsof a rnolecule
do not haveE :0: thevhavea "zero-point"energythat is approrimatelyequal
Lo I l2( h l21 11.1;
I p\ti: . [Link] actionquantization conditionyieldsen-
ergieswhosespacings arereasonablvin agreementrvith laboratorydatatbl lorv-
energystates(e.g..suchstateshaveapproxirnately constantspacings), it fails to
predictthe zero-pointenergycontentof [Link] we will seelirter,a
properquantumrrechanicaltreatmentof the harmonicoscillatoryieldsenc'rgies
of the ibrm
"
E : (,,.:) (*)(:) (1.143)

which differs from the oction-bascdresult by the proper zero-point energy.

 T h e b a s i c so f q u a n t u mm e c h a n i c s

Evenu'ith suchdifficultiesknou'[Link] hasbeenmadein extending

themostelementary action-based methodsto moreandmoresvstemsby introduc_
ing. for [Link] alloll the quantumnumbern to assumehalf-integer
a s ' " v e lal s i n t e - q e r v a l u eCsl .e a r l 1i,1. ' nu ' e r ea l l o w e dt o e q u a ll / 2 . 3 / 2 . 5 1 2 . . . . ,
the earlieractionintegralwouldhaveproducedthe [Link],
doesone know when to allorvr?to assumeonly integeror only harf-integer
or
both integerand half-integervalues?The [Link] to this questionare beyondthe
scopeofthis text and constitutean [Link] no$,.it is enough
for the studentto be awarethat one can often find energylevelsby,using
action
integrals. but onerrrustbe carefulin doingso because solnetlmes theanswersare
wrong.
Beforeleavingthis [Link] is worth noting that the appearance of half-
rntegerquantumnumbersdoesnot onlv occurin theharmonicoscillator -lb
case.
[Link] us considerthe r_- angularmomentumoperatordiscussed
earlier.
As we showed this [Link] [Link] as the ---componentof r x p, can
b e w r i t t e ni n p o l a r( r . 0 . 6 ) c o o r d i n a t eass

L_ : -ihd /d0. (l.l4l)

The eigenfunctions ofthis operatorha'e the forrr exp(ia@).andthe eigenvalues
arenfi. Becausegeometries with azimuthalanglesequal1o@or equalto * 2r
Q
areexactl)'the samegeometries, thefu'ction exp(ia(t)shouldbe exactlythesame
as exp(ia(Q't hr 1).This can onrybe the caseif a is an [Link].
one con-
cludesthatonlyintegralmultiplesoffi canbe'.allowed"valuesofthe:-cornponent
of angularmomenturn. Experimentalry,,
onemeasures the:-componentof an an-
-eularmomentum bi, placingthe systempossessing the a'gular momentumin a
magneticfield of strengthB and obser'ing how many---corxponent energystates
[Link],a boronatomwith its2p orbitalhasoneunit of orbital
angular
[Link] one finds three separate;-co,lponent 'alues r.l,hichare
usually
denotednt : -7, m :0, andnt = L Anotherexampleis offeredby the
scan_
dium atom rvith one unpairedelectronin a d orbital: this atom,sstatessplit
into
five (rr = -2, -1. 0. 1, 2) z--component [Link] [Link] finds ZL + l
valuesof the r/r quantumnumber,and,becauser is a' integer.2L
+ l is an odd
[Link] of theseobservationsare consistentwith the expectation
that onlv
integervaluescan occur for Z, eigenvalues.
[Link] hasbeenobservedthat somespeciesdo not possess3 or
5 or 7 or
9:-component statesbut an evennumberofsuch [Link] particular.
whenelec-
trons,protons,or neutronsaresubjectedto the kind of magneticfield
experiment
mentionedabo'e, theseparticlesare observedto have only two
z-componenr
eigenvalues. Because,as we discusslater in this text. all an-qularmomentahave
.?-componenteigenvaluesthat are separatedfrom one another by
unit muiti_
ples of fi. one is forced to concludethat thesethreefundamentalbuilding-block
particleshave:-componenteigenvalues of u2ri and.-r /2h. The appearance of
 F r e ep a r t i c l em o t i o n si n m o r e d i m e n s i o n s 45

half-integer angularmomentais not consistent with the observation madeearlier

that @and Q + 2;r correspondto exactlythe samephysicalpoint in coordinate
space, u'hich,in turn,impliesthatonly full-integerangularmomentaarepossible.
The resolutionof the aboveparadox(i.e., horv can half-integerangularmo-
mentaexist?)involvedrealizingthat someangularmomentacorrespondnot to
the r x p angularmomentaof a [Link]. insteadareintrinsic
propertiesof [Link] is, the intrinsicangularmomentaof elec-
trons,protons,and neutronscan not be viewed as arising from rotationof some
massthat [Link] intrinsic angularmomentaare
"built in"
fundamental characteristicsof [Link] [Link] trvo
I l2l1 and-112h angularmomentumstatesof an electron,usuallydenotedo and
p. respectively. are two internal statesofthe electronthat are degeneratein the
absenceof a magneticfield but which representfwo distinct statesof the elec-
tron. Analogousl,v, a proton has I l2h and -l l2li states,as do [Link] such
half-integerangularmomentumstatescan not be accountedfor using classical
mechanicsbut are knorvnto arisein quantummechanics.