2

Electron, Ion and Atom Scattering

M. R. Flannery October 30, 2000
School of Physics, Georgia Institute of Technology, Atlanta, Georgia 30332-0430

2.1 COLLISIONS 2
2.1.1 Differential and Integral Cross Sections 2
2.1.2 Collision Rates, Collision Frequency and Path Length 3
2.1.3 Energy and Angular Momentum: Center of Mass and Relative 3
2.1.4 Elastic Scattering 4
2.1.5 Inelastic Scattering 4
2.1.6 Reactive Scattering 4
2.1.7 Center-of-Mass to Laboratory Cross Section Conversion 4
2.2 MACROSCOPIC RATE COEFFICIENTS 5
2.2.1 Scattering Rate 5
2.2.2 Energy Transfer Rate 5
2.2.3 Transport Cross Sections and Collision Integrals 6
2.3 QUANTAL TRANSITION RATES AND CROSS SECTIONS 6
2.3.1 Microscopic Rate of Transitions 6
2.3.2 Detailed Balance Between Rates 7
2.3.3 Energy Density of Continuum States 7
2.3.4 Inelastic Cross Sections 8
2.3.5 Detailed Balance between Cross Sections 8
2.3.6 Examples of Detailed Balance 9
2.3.7 Four useful expressions for the cross section 9
2.4 BORN CROSS SECTIONS 10
2.4.1 Fermi Golden Rules 11
2.4.2 Ion (electron) - Atom Collisions 11
2.4.3 Atom - Atom Collisions 12
2.4.4 Quantal and Classical Impulse Cross Sections 12
2.4.5 Atomic Form Factor and Generalized Oscillator Strength 12
2.4.6 Form Factors for Atomic Hydrogen 13
2.4.7 Rotational Excitation 13
2.4.8 List of Born Cross sections for Model Potentials 13
2.5 QUANTAL POTENTIAL SCATTERING 14
2.5.1 Partial Wave Expansion 14
2.5.2 Scattering Amplitudes 15
2.5.3 Integral cross sections 15
2.5.4 Differential cross sections 16
2.5.5 Optical Theorem 16
2.5.6 Levinson’s Theorem 17
2.5.7 Partial Wave Expansion for Transport cross sections 17
2.5.8 Born Phase Shifts 17
2.5.9 Coulomb Scattering 18
2 / Electron, Ion and Atom Scattering

2.6 COLLISIONS BETWEEN IDENTICAL PARTICLES 18

2.6.1 Fermion and Boson Scattering 18
2.6.2 Coulomb scattering of two identical particles 19
2.6.3 Scattering of Identical Atoms 19
2.7 QUANTAL INELASTIC HEAVY-PARTICLE COLLISIONS 20
2.7.1 Adiabatic Formulation (Kinetic Coupling Scheme) 20
2.7.2 Diabatic Formulation (Potential Coupling Scheme) 20
2.7.3 Inelastic scattering by a Central Field 21
2.7.4 Two State Treatment 22
2.7.5 Exact Resonance 22
2.7.6 Partial Wave Analysis 23
2.8 ELECTRON-ATOM INELASTIC COLLISIONS 23
2.8.1 Close Coupling Equations for electron-atom (ion) collisions 23
2.8.2 Close Coupling with Pseudostates and Correlation 24
2.8.3 The R-Matrix Method 25
2.8.4 Electron-Molecule Collisions 25
2.9 SEMICLASSICAL INELASTIC SCATTERING 25
2.9.1 Classical Path Theory 25
2.9.2 Landau - Zener Cross Section 26
2.9.3 Eikonal Theories 26
2.10 LONG RANGE INTERACTIONS 27
2.10.1 Polarization, Electrostatic and Dispersion Interactions 27
2.11 FURTHER READING 29
2.12 LIST OF KEYWORDS 29

This Chapter deals with quantal and semiclassical section for scattering of the test particles into unit solid
theory of heavy-particle and electron-atom collisions. angle dΩ = d(cos ψ) dφ about the direction v̂
(ψ, φ) of
Basic and useful formulae for cross sections, rates and the final relative motion is
associated quantities are presented. A consistent de- number of test particles scattered by one field
scription of the mathematics and vocabulary of scatter- dσ(v; ψ, φ) particle into unit solid angle per unit time
=
ing is provided. Topics covered include Collisions, Rate dΩ current ji of incident beam

Coefficients, Quantal Transition Rates and Cross Sec- The number of particles scattered per unit time by the
tions, Born Cross Sections, Quantal Potential Scattering, field particle and detected per unit time is then
Collisions between Identical Particles, Quantal Inelastic
dNd dσ dσ dσ dA
Heavy-Particle Collisions, Electron-atom inelastic colli- = ji dΩ = NA v dΩ = NA v
sions, Semiclassical Inelastic Scattering and Long Range dt dΩ dΩ dΩ r 2
Interactions. where the detector, located along the scattered direction
v̂
(ψ, φ), subtends an angle dΩ = dA/r 2 at the scattering
center and projects an area dA = r 2 d(cos θ)dφ normal
2.1 COLLISIONS to the scattered beam. Thus [dσ/dΩ] dΩ is the cross
sectional area of the beam that is intercepted by one
2.1.1 Differential and Integral Cross target particle and scattered into the solid angle dA/r 2
Sections of a cone with axis along v̂
(ψ, φ) and vertical angle dψ.
In classical terms (Fig. 2.1), the number of particles
A uniform monoenergetic beam of test or projectile detected per second about direction (ψ, φ) is the number
particles A with number density NA and velocity vA is NA v (bdb dφ) of incident particles crossing the initial
incident on a single field or target particle B of velocity areal element bdb dφ per second. Hence
vB . The direction of the relative velocity v = vA − vB is dσ bdb
along the Z -axis of a Cartesian XYZ- frame of reference. = .
dΩ d(cos ψ)
The incident current (or intensity) is then ji = NA v,
which is the number of test particles crossing unit area For an incident current flowing between two cylinders
normal to the beam in unit time. The differential cross of radii b and b + db, then ji 2π[dσ/dΩ] d(cos ψ) is the
 Encyclopedia of Chemical Physics and Physical Chemistry / 3

= −k(E)NA (E)NB
= −νB (E)NA (E) .

in unit time and unit volume. The microscopic rate

coefficient (cm3 s−1 ) for the scattering of 1 test particle
by 1 field particle is k(E) = vσ(E). The f requency (s−1 )
of collision between 1 test particle and NB field particles
(cm−3 ) is νB = k(E)NB . Since v = dx/dt, the variation
with x of intensity ji of the attenuated beam is governed
by
dji
= −[NB σ(E)]ji (x)
dx
For constant density NB , the solution is

ji (E, x) = ji (E, 0) exp [−NB σ(E)x]

= ji (E, 0) exp(−x/λ)

Figure 2.1. Scattering of a beam with current ji = NA v where λ ≡ 1/NB σ(E) = v/νB is the path length
particles per unit area incident between two cylinders of between collisions. Since ji = NA v, the density NA (E, x)
radii b and b+db by one particle at rest in the Laboratory. obeys a similar equation. These equations describe the
attenuation of a particle beam A travelling through a
target gas B. For target gas particles with a distribution
number of particles scattered per second between the two fB (vB ) dvB in velocities vA , the above microscopic rate
cones of semivertical angles ψ, ψ + dψ (Fig. 2.1). then becomes
The integral cross section for scattering over all

directions is k(E) = |vA − vB |σ(|vA − vB |) fB (vB ) dvB

+1
2π 
dσ
σ(v) = d(cos ψ) (v; ψ, φ) dφ . where E = 12 MA vA2
is the kinetic energy of the projectile
dΩ beam. For an isothermal beam with an energy distribu-
−1 0
tion fA (E) dE at temperature T, the macroscopic rate
The integral cross section is therefore the effective coefficient (cm3 s−1 ) or thermal rate constant is
area presented by each field particle B for scattering of
∞
the test particles A into all directions. The probability k(T ) = k(E) fA (E) dE .
that the test particles are scattered into a given direction 0
v̂
(ψ, φ) is the ratio

dσ(v; ψ, φ) 2.1.3 Energy and Angular Momentum:

P(v; ψ, φ) = /σ(v)
dΩ Center of Mass and Relative
of the differential to integral cross sections.
The velocity of the center of mass of the projectile
and target particles of respective masses MA and MB is
2.1.2 Collision Rates, Collision
Frequency and Path Length V = (MA vA + MB vB )/(MA + MB ) .

An electron or atomic beam of (projectile or test) The relative velocity is

particles A with density NA of particles per cm3 travels v = vA − vB .
with speed v and energy E through an infinitesimal
thickness dx of (target or field) gas particles B at rest with The velocities of A and B in terms of V and v are
density NB particles per cm3 . The particles are scattered
out of the beam by (A − B) collisions with integral cross MB
vA = V + v
section σ(E) at a rate (cm−3 sec−1 ) given by the total MA + MB
number of collisions between A and B particles MA
vB = V − v.
MA + MB
dNA (E)
= − [NA (E)vσ(E)] NB The total kinetic energy then decomposes into the sum
dt
4 / Electron, Ion and Atom Scattering

2 2
E = 12 MA vA + 12 MB vB = 12 M V 2 + 12 MAB v 2 Changes in angular momentum are limited by the con-
= Ecm (A + B) + Erel (AB), servation requirement that

of the energy Ecm = 12 M V 2 of the CM with mass Lrel (i) + Lα (A) + Lβ (B) = Lrel (f ) + Lα
(A) + Lβ
(B)
M = (MA + MB ), and the energy Erel = 12 MAB v 2
of relative motion, where the reduced mass MAB is where Li denotes the internal angular momentum of
MA MB /(MA + MB ). Let R be the position of the each isolated species. Collisions, in which only angular
CM relative to a fixed origin O and r be the inter- momentum is transferred without any energy change, are
particle separation. The total angular momentum about called quasi-elastic collisions.
O similarly decomposes into the sum
2.1.6 Reactive Scattering
L = R × M V + r × MAB v
= Lcm (A + B) + Lrel (AB) A+B →C +D

of angular momenta of the CM and of relative motion. Reactive Scattering or a Chemical Reaction is character-
For any collision in the absence of any external field, ized by a rearrangment of the component particles within
the energy Ecm and angular momentum Lcm of the CM the collision system, thereby resulting in a change of
are always conserved for all types of collision. The physical and chemical identity of the original collision re-
two species, A and B , may be electrons, ions, atoms actants A + B into different collision products C + D.
or molecules, with or without any internal structure Total mass is conserved. The reaction is exothermic
and may therefore possess internal energy and angular when Erel (CD) > Erel (AB) and is endothermic when
momentum which must be taken into account. For Erel (CD) < Erel (AB). A threshold energy is required
structured particles Erel and Lrel can change in a for the endothermic reaction.
collision.
2.1.7 Center-of-Mass to Laboratory
2.1.4 Elastic Scattering Cross Section Conversion
A(α) + B(β) → A(α) + B(β) Theorists calculate cross sections in the Center-of-
Mass frame while experimentalists usually measure cross
Elastic Scattering involves no permanent changes in the
sections in the Laboratory frame of reference. The
internal structures (states α and β) of A and B . Both the
Laboratory (Lab) System is the coordinate frame in
energy Erel and angular momentum Lrel (AB) of relative
which the target particle B is at rest before the collision
motion are therefore all conserved.
i.e., vB = 0. The Center of Mass (CM) System ( or
Barycentric System ) is the coordinate frame in which
2.1.5 Inelastic Scattering the CM is at rest i.e., V = 0. Since each scattering
of projectile A into (ψ, φ) is accompanied by a recoil of
A(α) + B(β) → A(α
) + B(β
) target B into (π − ψ, φ + π) in the CM -frame, the cross
sections for scattering of A and B are related by
Inelastic Scattering produces a permanent change in the      
internal energy and angular momentum state of one or dσ(ψ, φ) dσA (ψ, φ) dσB (π − ψ, φ + π)
both structured collision partners A and B, which retain ≡ =
dΩ dΩ dΩ
their original identity after the collision. For inelastic i ≡
CM
 CM
  CM
dσB (ψ, φ) dσ(π − ψ, φ − π)
(α, β) → f ≡ (α
, β
) collisional transitions, the energy = .
dΩ dΩ
Ei,f = 12 MAB vi,f2
of relative motion, before (i) and CM CM

after (f ) the collision satisfies the energy conservation In the Lab frame, the projectile is scattered by θA and
condition, the target, originally at rest, recoils through angle θB .
The number of particles scattered into each solid angle
Ei + &α (A) + &β (B) = Ef + &α
(A) + &β
(B)
in each frame remains the same, the relative speed v is
where &A,B are the internal energies of A and B. The max- now vA and ji = NA v in each frame. Hence
imum amount of kinetic energy that can be transferred     
dσA (θA , φ) dΩ dσ(ψ, φ)
to internal energy is limited to the initial kinetic energy = ,
dΩA dΩA dΩ
of relative motion, Erel (AB) = 12 MAB vi2 . Excitation im-  Lab   CM
plies implies &i ≡ &α (A) + &β (B) > &α
(A) + &β
(B) ≡ &f ., dσB (θB , φ) dΩ dσ(ψ, φ)
= .
de-excitation (or superelastic) implies &f < &i and en- dΩB Lab dΩB dΩ CM
ergy resonance or excitation- transfer implies &i = &f .
 Encyclopedia of Chemical Physics and Physical Chemistry / 5

Two-body elastic scattering Collisions with B will scatter A out of the beam at the
loss rate (cm−3 s−1 )
A(α) + B(β) → A(α) + B(β)
dNA
= −kNA (t)NB (t) = −νB NA (t) .
The scattering and recoil angles θA and θB in the Lab dt
frame are related to the CM scattering angle ψ by
The macroscopic rate coefficient k (cm3 s−1 ) for elastic
sin ψ collisions between the ensembles A and B is
tan θA = , γ = MA /MB


1 + γ cos ψ
3 −1
θB = 12 (π − ψ) 0 ≤ θB ≤ 12 π . k(cm s ) = fA (vA ) dvA [v σ(v)] fB (vB ) dvB

The elastic cross sections for scattering and recoil in the in terms of the integral cross section σ(v) for A − B
Lab-frame are related to the cross section in the CM- elastic scattering at relative speed v = |vA − vB |. The
frame by microscopic rate coefficient is vσ(v). The frequency νB
    (s−1 ) of collision between one test particle A with NB
2 3/2
dσA (θA , φ) (1 + γ + 2γ cos ψ) dσ(ψ, φ)
= , field particles is kNB .
dΩA |1 + γ cos ψ| dΩ The rate coefficient for elastic scattering between two
 Lab   CM
dσB (θB , φ)   dσ(ψ, φ) species with non-isothermal Maxwellian distributions is
= 4 sin 12 ψ  . then
dΩB Lab dΩ CM

∞
k(cm3 s−1 ) = ṽAB σ(˜
&rel ) &˜rel exp (−˜
&rel ) d˜
&rel ,
0
Two-body inelastic or reactive scattering
where
process A + B → C + D
  1/2
8kB TA TB
The energies Ei and Ef of relative motion of A and B ṽAB = +
π MA MB
and of C and D, respectively satisfy Ef /Ei = 1 − &f i /Ei ,
where &f i = &f − &i is the increase in internal energy. The and
scattering and recoil angles are
1 MA MB v 2
 1/2  1/2 &˜rel =
2 kB (MA TB + MB TA )
.
sin ψ MA MC Ei
tan θC = , γC =
(γC + cos ψ) MB MD Ef For isothermal distributions TA = TB = T , the rate is
 1/2  1/2
sin ψ MA MD Ei
∞
tan θD = , γD = −
(|γD | − cos ψ) MB MC Ef k(T ) = vAB σ(&) & exp (−&) d& (cm3 s−1 ),
0
The Lab and CM cross sections are then related by 1 2
where & = 2 MAB v /kB T and vAB =
  2 3/2   (8kB T /πM AB ) 1/2
. The rate of collisions of electrons
dσj (θj , φ) 1 + 2γj cos ψ + γj dσ(ψ, φ)
= . A at temperature Te with a gas of heavy particles B at
dΩj |1 + γj cos ψ| dΩ CM
Lab temperature TB is
where j denotes C or D. The scattering of a beam from a
∞
stationary target are governed by the above equations. A k(Te ) = ve σ(&e ) &e exp (−&e ) d&e (cm3 s−1 ),
0
crossed beam experiment in which two beams intersect
at an angle is not in the Lab frame. In this case the where &e = 12 me v 2 /kB Te and ve = (8kB Te /πme )1/2 .
measured quantities can be similarly transformed [1] to
CM for comparison with theoretical calculations.
2.2.2 Energy Transfer Rate
2.2 MACROSCOPIC RATE Each of the species A transfers energy EAB to each
COEFFICIENTS species B. The amount of energy transferred per unit
volume in unit time from ensemble A to ensemble B is
2.2.1 Scattering Rate d
[NA EAB ] = −kE NA (t)NB (t) = −νEB NA (t)
dt
A distribution fA (vA ) of NA (t) test particles (cm−3 )
of species A in a beam collisionally interacts with a where the macroscopic rate coefficient kE (energy cm3
distribution fB (vB ) of NB (t) field particles of species B. s−1 ) for the averaged energy loss EAB is
6 / Electron, Ion and Atom Scattering



kE = fA (vA ) dvA fB (vB ) dvB collision so that the ions move with a constant drift speed
  vd = KE. The mobility K of ions of charge e in a gas


dσ of density N is given in terms of the collision integral by
EAB (vA , vB ; ψ, φ) v dΩ .
dΩ the Chapman-Enskog Formula[2]
 1/2
The amount of energy lost in unit time, the energy- 3e π −1
loss frequency, is νEB = kE NB (t). The energy-loss rate K= Ω(1,1) (T ) .
16N 2M kB T
coefficient for two-temperature Maxwellian distributions
is
2MA MB 2.3 QUANTAL TRANSITION RATES
kE (TA , TB ) = kB (TA − TB )ṽAB
(MA + MB )2 AND CROSS SECTIONS

∞
× σD (˜ &rel )2 exp(−˜
&rel ) (˜ &rel ) d˜
&rel 2.3.1 Microscopic Rate of Transitions
0

where σD (˜
&rel ) is the momentum transfer cross section In the general elastic/inelastic collision process
at reduced energy &˜rel . For isothermal distributions,
TA = TB and the energy rate coefficient kE of course A(α) + B(β) → A(α
) + B(β
)
then vanishes. the external scattering or deflection of a beam of projec-
tile particles A (electrons, ions, atoms) by target par-
2.2.3 Transport Cross Sections and ticles B (atoms, molecules) is accompanied by transi-
tions (electronic, vibrational, rotational) within the in-
Collision Integrals
ternal structure of either or both collision partners. For
Transport cross sections are defined for n ≥ 1 as a beam with incident momentum pi = h̄ki in the range
(pi , pi + dpi ) or directed energy E 1 i ≡ (Ei , p̂i ) in the
 −1 1 i, E
1 i + dE1 i ), the translational states represent-
1 + (−1)n range (E
σ (n) (E) = 2π 1 − ing the A − B relative or external motion undergo free-
2(n + 1)

+1 free transitions (E 1 i, E1 i + dE
1 i ) → (E
1f , E
1 f + dE
1 f ) within
dσ
× [1 − cosn θ] d(cos θ) . the translational continuum, while the structured parti-
−1 dΩ cles undergo bound-bound (excitation, de-excitation, ex-
The diffusion and viscosity cross sections are given by the citation transfer) or bound-free (ionization, dissociation)
transport cross sections σ (1) and 23 σ (2) , respectively. transitions i ≡ (α, β) → f ≡ (α
, β
) in their internal
Collision Integrals are defined as electronic, vibrational or rotational structure. The tran-
sition frequency (s−1 ) for the above collision is,
−1
Ω(n,s) (T ) = (s + 1)!(kB T )s+2 dWif −1 2π 1

∞ (s ) = |Vf i |2 ρf (Ef )
σ (n) (E) E s+1 exp(−E/kB T )dE dp̂f h̄ gi
i,f
0

∞
which is an average over the gi initial degenerate internal
= [(s + 1)!]−1 σ (n) (&) &s+1 exp −& d& states and a sum over all gf final degenerate internal
0
states of the isolated systems A and B. It is therefore the
1 2
where & = 2 MAB v /kB T . The external factors are probability per unit time for scattering from a specified
chosen so that the above expressions for σ (n) and Ω(n,s) E1 i - (external) continuum state into unit solid angle
reduce to πd2 for classical rigid spheres of diameter dp̂f accompanied by a transition from any one of the
d. The rate coefficient k (cm3 s−1 ) for scattering can gi - initial states (α, β) to all final internal states (α
, β
)
then be expressed, in terms of the collision integral, as of degeneracy gf and to all final translational states
= ṽAB Ω(0,0) . The amount of energy lost per cm3 per ρf (Ef ) dEf of relative motion consistent  with energy
second by collision can be expressed in terms of Ω(1,1) . conservation. The double summation i,f is over the gi
Tables of transport cross sections and collision integrals initial and gf final internal states of A and B with total
for (n,6,4) ion -neutral interactions are available[2, 3]. energy &i and &f respectively.
Check:The dimension of [|Vij |2 ρ] is E, [h̄] = Et so
that dWif /dp̂f indeed has the correct dimension of t−1 .
Chapman-Enskog Mobility Formula
When ions move under equilibrium conditions in a gas Interaction Matrix Element
and an external electric field, the energy gained from the
electric field E between collisions is lost to the gas upon The matrix element
 Encyclopedia of Chemical Physics and Physical Chemistry / 7

Vf i = Nf Φf |V (rA , rB , R)|Ni Ψ+
i r,R = Vif∗ ρi dEi initial external translational states is equal to the
reverse frequency from the gf degenerate final internal
is an integration over the internal coordinates r ≡ rA , rB states and the ρf dEf final external translational states.
of the electrons of A and B and over the channel vector The Detailed Balance relation between the forward and
R for A − B relative motion. The matrix element of reverse frequencies is therefore
the mutual electrostatic interaction V (rA , rB , R) couples    
the eigenfunction Ni Ψ+ i (R, rA , rB ) of [Ĥrel + Ĥint + [gi ρi dEi dp̂i ]
dWif
dp̂f = [gf ρf dEf dp̂f ]
dWf i
dp̂i
V ] for the complete collision system for all R to the dp̂f dp̂i
final R → ∞ asymptotic state Nf Φf (R, rA , rB ), which
is an eigenfunction only of the unperturbed Hamiltonian since Vif = Vf∗i . From energy conservation &i + Ei =
[Ĥrel + Ĥint ]. The wavefunction &f + Ef , then dEi = dEf . The differential frequencies
dRif dWif dWf i dRf i
Ψ+
i (R, rA , rB ) = Φ+
j (R)ψα (rA )φβ (rB )
≡ gi ρi = gf ρf ≡
dp̂f dp̂f dp̂i dp̂i
j

≡ Φ+ int for the forward and reverse transitions, i 9

: f , are
j (R)ψj (r)
j
therefore equal.

for the full collision system with Hamiltonian Ĥrel + 2.3.3 Energy Density of Continuum
Ĥint + V tends at asymptotic R to
States
 
ikj ·R eikj R
Ψi ∼
+
e δij + fij (θ, φ) ψjint (r) The continuum wavefunctions φp (R) for the states
j
R
of the A-B relative motion satisfy the orthonormality
which represents an incoming plane wave of unit ampli- condition



tude in the incident elastic channel i and an outgoing
ρ(E) dE φp (R)φ∗p
(R) dR = 1 .
spherical waves of amplitude fij in all channels j, includ-
ing i. The Kroneker symbol means δij = 1, i = j and
δij = 0, i = j. The final state at infinite separation R is The number of translational states per unit volume dR
with directed energies E ≡ (E, p̂) in the range [E, E+dE]
Φf (R, rA , rB ) = eikf ·R ψα
(rA )φβ
(rB ) ≡ eikf ·R ψfint (r) is ρ(E) dE. This orthonormality condition
 for continuum
states is analogous to the condition j | φj |φi |2 = 1 for
which is an eigenfunction only of Ĥrel + Ĥint . The plane bound states. For plane waves, φp (R) = N exp(ip·R/h̄),
wave of unit amplitude describes the external relative then
motion with Hamiltonian Ĥrel and φα
(rA )ψβ
(rB ) de-
scribes the internal isolated atomic states of A and B φp
|φp = |N |2 (2πh̄)3 δ(p − p
) = |N |2 (2π)3 δ(k − k
)
with internal Hamiltonian Ĥint . The factors Ni,f pro- (2πh̄)3 1
= |N |2 δ(E − E
) = δ(E − E
) .
vide the possibility of having translational (scattering) mp ρ(E)
states with arbitrary amplitudes which are not necessar-
ily unity. Note, irrespective of the method chosen to normalize the
wavefunctions, that
mp
Transition Operator |N |2 ρ(E) =
(2πh̄)3
The interaction matrix element can also be written as always. The amplitude |N | does however depend on the
choice of normalization.
Vf i = Nf Φf |T̂ |Ni Φi .
1. For momentum normalized states, φp
|φp = δ(p −
where the transition operator, T̂ , is defined by T̂ Φ = V Ψ. p
), |N | = (2πh̄)−3/2 and the density of states ρ(E) =
The transition operator T̂ therefore couples states which mp.
are eigenfunctions of the same unperturbed Hamiltonian
2. For wavevector normalized states, φp
|φp = δ(k −
Ĥrel + Ĥint , in contrast to V which couples states Ψ+ i
k
), |N | = (2π)−3/2 and the density of states ρ(E) =
and Φf belonging to different Hamiltonians.
(mp/h̄3 ).
3. For energy-normalized states, φp
|φp = δ(E − E
),
2.3.2 Detailed Balance Between Rates ρ(E) = 1 and |N | = (mp/h3 )1/2 .
The frequency (number per sec.) of i → f transitions 4. For waves with unit amplitude, |N | = 1 and ρ(E) =
from all gi degenerate initial internal states and from the (mp/h3 ).
8 / Electron, Ion and Atom Scattering

Note that φp
|φp ρ(E) dE is dimensionless for all Basic Expression for Cross Section
cases and yields unity for a single particle when inte-
grated over all E. The number of states in the phase- The differential cross section for
space element dEdR is
A+B →C +D
dn = |N |2 ρ(E) dE dR = dp dR/(2πh̄)3 .
collisions is therefore defined as
ie., each translational state occupies a cell of phase
volume (2πh̄)3 . The density of states in the interval dσif 1 dRif
=
[E, E + dE] is ρ(E) = 4πρ(E). The number of trans- dp̂f gi ji dp̂f
 
lational states per unit volume with energy in the scalar 2π ρi ρf 1
range [E, E+dE] is = | Nf Φf |V (rA , rB , R)|Ni Ψ+
i |
2
h̄ ji gi
i,f
2 (2πm)3/2 1/2
|N |2 ρ(E) dE = √ E dE . which is an average over the gi initial internal degenerate
π h3
states and a sum over the gf final degenerate states. The
Check: The number of free particles with all momenta external multiplicative factors γif = (2π/h̄) (ρi ρj /ji ) ap-
p in equilibrium with a gas bath of volume V at propriate to various normalization schemes are summa-
temperature T is the translational partition function rized in Table ??.?.?. Since ji = |Ni |2 vi ρi = p2i /h3 , an
Zt . Since the fraction of particles with energy E is alternative form [4] for the cross section is
exp(−E/kB T )/Zt , the Maxwell distribution  
dσif 2π ρf
|N |2 V ρ(E) dE = | Nf Φf |V (rA , rB , R)|Ψ+
i | .
2

fM (E)dE = exp −(E/kB T ) dp̂f h̄vi gi

i,f
Zt
2
= √ (E/kB T )1/2 exp(−E/kB T ) d(E/kB T )
π
2.3.5 Detailed Balance between Cross
is then recovered. Sections

Current When cast in terms of cross sections, the detailed

balance relation in Section 2.3.2 is
Current is the number of particles crossing unit area
1 i ) dσif = gf jf (E
gi ji (E 1 f ) dσf i .
in unit time. The current in a beam with directed energy dp̂f dp̂i
E within the range (E, E + dE) is
The basic relationship satisfied by the differential cross
j dE = v|N |2 ρ(E) dE = (p2 /h3 ) dE .
sections for the forward and reverse i 9
: f transitions is
The current per unit dE is the current density j =
dσif (Ei ) dσf i (Ef )
(p2 /h3 ). The quantal expression for current gi p2i = gf p2f .
dp̂f dp̂i
h̄
J≡ [φ∗ ∇φp − φp ∇φ∗p ] ,
2mi p
when applied to the plane wave φp = N exp(ip · R/h̄), Collision Strengths
gives J = |N |2 v. The current in a (E, E + dE)-beam
of plane waves is then J[ρ(E) dE] so that the current Collision Strengths Ωif exploit this detailed balance
density is j(E) = Jρ(E) = |N |2 v ρ(E), as before. relation by being defined as

Ωif = gi p2i σif (Ei ) = gf p2f σf i (Ef ) = Ωf i .

2.3.4 Inelastic Cross Sections
They therefore symmetrical in i and f .
The differential cross Section dσif /dp̂f for i → f
transitions from any one of the gi initial states is defined
Reactive Processes
as [dRif /dp̂f ]/gi ji , the transition rate per unit incident
current. Since current is the number of particles crossing
For any reactive process
unit area in unit time, the cross section is therefore
the effective area presented by the target towards i → A+B 9
:C+D
f internal transitions in the internal structures of the
collision partners which are scattered into unit solid angle the detailed balance relations involving differen-
dp̂f about direction p̂f in the CM-frame. tial/integral cross sections are
 Encyclopedia of Chemical Physics and Physical Chemistry / 9
   
dσif (EAB ) dσf i (ECD ) photon energy E is related to its momentum pν and
gA gB p2AB = gC gD p2CD
dp̂CD dp̂AB wavenumber kν and to the ionization energy In" of the
gA gB p2AB σif (EAB ) = gC gD p2CD σf i (ECD ) atom A(n<) by

where p2JK = 2MJK EJK , in terms of the reduced mass E = hν = pν c = h̄kν c = In" + Ee
MJK and relative energy EJK of species J and K. where c is the speed of light. The ratio above is
p2ν (hν)2 α2 (hν)2
2.3.6 Examples of Detailed Balance 2
= 2
= .
pe (2Ee me c ) 2Ee ε0
Excitation - De-excitation
e− (Ei ) + Ai 9
: e− (Ef ) + Af 2.3.7 Four useful expressions for the
cross section
  
gf Ef Final expressions to be used for calculation of cross
σif (Ei ) = σf i (Ef )
gi Ei sections depend on the particular choice of normalization
of the continuum wavefunction for relative motion. Since
With energy conservation, Ei = Ef + (&f − &i ) ≡ Ef + &f i
it is often a vexing problem and is a continued source of
the cross section for superelastic collisions (Ef > Ei ) can
confusion and error in the literature, these final expres-
be obtained from σif at energy (Ei ) by
sions are worked out below. The external relative-motion
 −1   part of the system wavefunction N Ψ+ ≡ Ψp (R, rA , rB )
&f i gi
σf i (Ei − &f i ) = 1− σif (Ei ) . is N φp (R). Since |N |2 ρ = mp/h3 , the density ρ(E) of
Ei gf continuum states therefore depends on the choice of nor-
malization factor N adopted for the continuum wave. For
future reference, the amplitude N and the energy densi-
Dissociative Recombination /Associative ties ρ(E) associated with four common methods adopted
Ionization for normalization of continuum waves are summarized in
Table ??.?.?. Also included are the amplitude N" of the
Dissociative Recombination and Associative Ioniza- corresponding radial partial wave
tion are represented by the forward and backward direc-
N" 1
tions of R#" (r) ∼ sin(kr − <π + η" )
r 2
e− + AB + 9
: A + B∗ . of Section (2.5.1). The external multiplicative factors
γif = (2π/h̄) (ρi ρj /ji ) in the basic formula in Section
The respective cross sections σDR and σAI are related by
(2.3.4) for the cross section are also summarized in
  2 
gA gB kAB Table ??.?.? for the various normalization schemes.
σDR (Ee ) = σAI (EAB ).
ge gAB + ke2
where the statistical weight of each species j involved is
Energy Normalized initial and final states
denoted by gj .
The wavefunctions
1 = ρ1/2 N Ψ = (mp/2πh̄3 )1/2 Ψk (R, rA , rB )
χ(E)
Radiative Recombination /Photoionization
are energy − normalized according to
Similarly, the cross sections for Radiative Recombi-
nation and for Photoionization, the forward and reverse χ(E
)|χ(E) = δ(E − E
) .
directions of
The basic formula in section 2.3.4 with ji = p2i /(2πh̄)3 ,
+ −
A +e 9
: hν + A(n<) yields
 
are related by dσif π 1 2 h2 1 2
= 2 |Tif | = |Tif | .
   dp̂f ki gi 8πmi Ei gi
gA gν p2ν
σRR (Ee ) = σP I (hν) The transition probability is
ge gA+ p2e


 
The photon statistical weight is gν = 2, corresponding to Pif = |Tif |2 = 2π χ̃f | V | χ+ 2 dk̂i ,
i
the two directions of polarization of the photon. The i,f
10 / Electron, Ion and Atom Scattering

the magnitude squared of the element Tif of the tran- Momentum Normalized initial and final states
sition matrix T between χ+ i and χ̃f , the two energy-
normalized eigenfunctions of Ĥrel + Ĥint + V and Ĥrel + Here the initial and final wavefunctions ξp =
Ĥint , respectively. The detailed balance relation in this (2πh̄)−3/2 Ψ+ ˜
i and ξp
= (2πh̄)
−3/2
Φf are normalized ac-
case is simply, cording to

|Tif |2 = |Tf i |2 ξk
|ξk = δ(p − p
) .

2
thereby verifying that |Tif | is indeed the i → The cross section 2.3.4 is then
f transition probability for transitions between
all gi initial and gf final states. This type dσif vf
= |fif (θ, ϕ)|2
of normalization is convenient for Rearrangement dp̂f vi
Collisions such as Dissociative, Radiative and
Dielectronic Recombination. where the scattering amplitude [6–8] is now

fif = −(2π 2 h̄3 )(2mf /h̄2 )| ξ˜f |V |ξi+ | .

Unit Amplitude initial and final states
Here the initial and final wavefunctions with unit
Energy Normalized final and unit amplitude
amplitude are Ψ+i and Φf . They are each normalized
according to initial states

Ψk
|Ψk = (2π)3 δ(k − k
) . Here the basic formula 2.3.4 yields

The basic expression in section 2.3.4 with ji = |Ni |2 vi ρi dσif 2π  2


= χ̃f | V | Φ+
and |Nf |2 ρf = mf pf /h3 reduces to dp̂f h̄vi i

dσif vf which couples the initial scattering state Ψ+

= |fif (θ, ϕ)|2 i of unit
dp̂f vi amplitude with the energy-normalized final state χ̃f =
1/2
where the scattering amplitude is ρf Nf Φf .This normalization is customary for photoion-
ization problems.
1 
fif = − (2mf /h̄2 ) Φf | V | Ψ+
i
4π
2.4 BORN CROSS SECTIONS
which couples scattering states Ψ+ i and Φf of unit am-
plitude. This expression is also applicable for rearrange- Here an (undistorted) plane wave of unit amplitude
ment collisions A + B → C + D by including the reduced is adopted for the channel wavefunction Φj (R) in the
mass mf = MC MD /(MC + MD ) of the reacted species wavefunction Ψ+ int
i = Σj Φj (R)ψj (r) for the complete
after the collision. The integral cross section consistent system. The differential cross section for elastic (i = f )
with the above scattering amplitude is, or inelastic scattering (i = f ) into k̂f (θ, φ) is then


2π
vf π 2
σif (E) = d(cos θ) |fif (θ, ϕ)| dϕ . dσif
=
vf
|fif (θ)|2 .
vi 0 0 dΩ vi
at relative energy E = ki2 h̄2 /2MAB . The scattering The Born scattering amplitude for A − B collisions is
amplitude consistent with the common use of



2π  2 (B)
fif (K) = −
1 2MAB
Vf i (R) exp(iK · R) dR
kf π ˜  4π h̄2
σif (E) = d(cos θ) fif (θ, ϕ) dϕ .
ki 0 0
which is the Fourier Transform of the interaction poten-
for rearrangement collisions, is tial
1 √   
f˜if = − (2 mi mf /h̄2 ) Φf | V | Ψ+
i . Vf i (R) = ψfint (r) | V (r, R) | ψiint (r)
4π
Both conventions are identical only for direct collisions which couples the initial and final isolated states
A(α)+B(β) → A(α
)+B(β
). This normalization is cus- ψjint (r) = φj (rA )ψj (rB ) of the atoms. The diagonal po-
tomary [5] for elastic and inelastic scattering processes. tential Vii (R) is the static interaction for elastic scatter-
For central potentials V (r) scattering is confined to a ing. The Born scattering amplitude is a pure function
plane and fij depends only on scattering angle θ = k̂i · k̂f . only of the collisional momentum change
 Encyclopedia of Chemical Physics and Physical Chemistry / 11

q = h̄K = MAB (vi − vf ) = h̄(ki − kf ) 2.4.2 Ion (electron) - Atom Collisions

2
where v is the AB relative velocity. Since K = ki2 +
The electrostatic interaction between a projectile ion
kf2 − 2ki kf cos θ, the Born integral cross section is
P of charge Zp e and an atom A with nuclear charge ZA e

q+  is
2π  (B) 2
B
σif (ki ) = 2 v2 fif (q) qdq A N
MAB i ZA Zp e2 Zp e2
q− V (r, R) = −
R |R − rj |

+ a0
K
 2
j=1
2π  (B) 
= fif (K) (Ka0 )d(Ka0 ) With the use of Bethe’s integral
(ki a0 )2
K− a0

eiK·R 4π
where q± = h̄K± = h̄|ki ± kf | are the maximum and dR = 2 eiK·rj
|R − rj | K
minimum momentum changes consistent with energy
conservation. For symmetric interactions Vf i (R), then the Born scattering amplitude (2.4) reduces to


2MAB sin KR 2  
fif (K) = −
B
Vf i (R) R dR .  (B) 2 4MP2 A Zp2 e4  2
h̄ 2 KR fif (q) = ZA δif − Fif
A
(q)
q 4

which is a function only of momentum transfer q = h̄K.

2.4.1 Fermi Golden Rules The dimensionless inelastic Form Factor for i → f
inelastic transitions between states φi,f of atom A with
Rule #2. The transition rate (probability per unit
ZA electrons is defined as
time for a transition) from state Φi of a quantum system
to a number ρ(1&) d1& of continuum states Φ# by an external ZA
perturbation V is FfAi (q) = φf (r) | eiq·rj /h̄ | φi (r) .
2π 2π j=1
wif = | Φ# | V | Φi |2 ρf (1&) ≡ |Vi# |2 ρf (1&)
h̄ h̄ where the integration is over all electron positions r ≡
to first order in V . Since |Vi# |2 ρf has dimension of energy {rj }. The integrated cross section is
&f , wif has dimension t−1 .
Rule #1. When the direct coupling Vi# from only the
q+
8πZp2 e4  2 dq
initial state to the continuum vanishes but the coupling σif (vP A ) = ZA δif − Fif
A
(q)
vP2 A q3
Vn# = 0 for n = i, the transition can then occur via the q−
intermediate states n at the rate
 
+ a0
K
 Vin Vn# 2 8πZp2 a20  2 d(Ka0 )
2π   = ZA δif − Fif
A
(K)
wif =  & − En  ρf (&) . (vP A /v0 )2 (Ka0 )3
h̄ n K− a0

These rules (which are not exact) are useful for both
since v0 = e2 /h̄. The dimensionless momentum-change
scattering and radiative processs.
q/me v0 is Ka0 . In the heavy particle or high energy
limit, q+ → ∞ and
Scattering Example  
|∆Ef i | ∆Ef i
q− ≈ 1+
The cross section for inelastic scattering of beam of vP A 2MP A vP2 A
particles by potential V (r, R) is
where ∆Ef i = Ef −Ei is the energy lost by the projectile.
dσif wif Since
= .
dp̂f Ji Z ZA eZ
B
fif (q) = fC p (q)δif + fC p (q)Fif (q)
A plane wave monoergic beam , Φi,f = N i, f exp(ipi,f ·
R/h̄)φi,f (r), has current Ji = |Ni |2 vi and density can be expressed in terms of the individual two-body
determined from |Nf |2 ρf (1&) = MAB pf /(2πh̄)3 . Hence amplitudes fC for Coulomb (a − b) elastic scattering, the

2 Born cross section for inelastic collisions can be written
dσ vf  1 2MAB 
=  2 Vf i (r)ei(pi −pf )·r/h̄
dr . [9, 28, 11] as
dp̂f vi 4π h̄
Since this agrees with the First Born differential cross
q+  
B 2π dσ
section for (in)elastic scattering, Fermi’s Rule #2 is σif (vi ) = 2 2 Pf i (q) qdq
MeP vi dΩ el
therefore valid to first order in the interaction V. q−
12 / Electron, Ion and Atom Scattering
 2

  ββ
cross section [6, 9] for general feB has recently been pre-
where Pf i (q) = FfAi (q) is the transition probability for
sented in a valuable new form [28] which is the appro-
which the impulsive transfer of momentum q to atom A
priate representation for direct classical correspondence.
and where
The classical impulse cross section was then defined [28]
  2
dσ 4MeP Zp2 e4 to yield the first general expression for the classical im-
= pulse cross section for n< − n
<
and n< − &<
electronic
dΩ el q4
transitions. The cross section satisfies the optical the-
is the cross section for elastic (Coulomb) scattering with orem and detailed balance. Direct connection with the
momentum q transferred from the projectile of charge classical binary encounter approximation (BEA) was es-
Zp e to a one electron of atom A. tablished and the derived n<−n
and n<−& cross sections
reproduce the standard BEA cross sections.
2.4.3 Atom - Atom Collisions
2.4.5 Atomic Form Factor and
The Born integral cross section for specific (αβ) → Generalized Oscillator Strength
(α
β
) transitions in the collision
In terms of the Form Factor Fif (K), the Generalized
A(α) + B(β) → A(α
) + B(β
) Oscillator Strength is defined as
 
in terms of the atomic Form Factors is 2me Ef i 2 2 Efa.u.
i
fif (K) = |F if (q)| = |Fif (K)|2
q2 (Ka0 )2

+ a0
K
ββ
8πa20  2 which tends to the dipole oscillator strength in the K → 0
σαα
(vi ) = ZA δαα
− Fαα
A

(K)
 ×
(vi /v0 )2 limit.
K− a0
 2 d(Ka0 )
ZB δββ
− Fββ
B

(K)
 Sum Rules
(Ka0 )3


∞
dfiE
fif (K) = fif (K) + dE = N ,
dE
f f 0
2.4.4 Quantal and Classical Impulse
Cross Sections

∞
dFiE
| Fif (K)| = 2
Fif (K) + dE
In the impulse approximation [6, 9], the integral cross dE
f f 0
section for i → f transitions in A is
N

q+  =N+ 1 · (1rj − 1rk ))|Ψi |2
| Ψi | exp(iK
2π  ββ
2  A 2
σif (vi ) = feB (q) Fif (q) q dq j<k
(MeB vi )2 

q− where N is the number of electrons. The summation
f
ββ
extends over all discrete and continuum states.
where feB is the scattering amplitude for elastic β =
β
or inelastic β = β
collisions between projectile B
and an orbital electron of A. For structureless ions
ββ
Energy change moments
B, the Coulomb feB (q) for elastic electron-ion collisions
reproduces the Born approximation for B-A collisions. The energy - change moments are defined as
ββ

When Born amplitudes feB (q) are used for fast atom
S(α, K) = (2∆Efa.u. α
i ) fif (K)
B − e collisions, then the Born approximation for atom-
f =i
atom collisions is also for general scattering amplitudes
|Fif (K)| (Ka0 )−2
2
recovered. For slow atoms B, feB is dominated by s-wave = (2∆Efa.u.
i )
α+1

elastic scattering and feB = −a and σeB = 4πa2 where f =i

a is the scattering length. Then The exact energy - change moments for H(1s) are

+ a0
K 1
S(−1, K) = {1 − [1 + (Ka0 )2 ]−4 }(Ka0 )−2 ,
2πa2  A 
σif (vi ) = Fif (K)2 (Ka0 ) d(Ka0 ) 4
(vi /v0 )2 S(0, K) = 1 ,
K− a0
S(1, K) = (Ka0 )2 + 4/3 ,
which is a good approximation for collisional transitions S(2, K) = (Ka0 )4 + 4(Ka0 )2 + 16/3 .
nl → n
l
in Rydberg atoms A. The full quantal impulse
 Encyclopedia of Chemical Physics and Physical Chemistry / 13

2.4.6 Form Factors for Atomic 2.4.7 Rotational Excitation

Hydrogen
For ion-point Dipole D Interactions, only ∆J = ±1
transitions are allowed. For ion-point Quadrupole Q In-
The probability of a transition i → f resulting from
teractions only ∆J = 0, ±2 transitions are allowed. The
any external perturbation which impulsively transfers

Born Differential Cross Sections for J → J transitions
momentum q to the internal momenta of the electrons
are
of the target system is
 
dσ (d) 4 kf J + 1 D2
Pif (q) = |Ff i (q)|2 . (J → J + 1) =
dk̂f 3 ki 2J + 1 K 2

The impulse can be due to sudden collision with particles

or to exposure to electromagnetic radiation. The physical dσ (q) 4 J(J + 1)
(J → J) = Q2
significance of the form factor is that Pif is the impulsive dk̂f 45 (2J − 1)(2J + 3)
transition probability for any atom. For nl → n
l

transitions in atomic hydrogen,
dσ (q) 2 kf (J + 1)(J + 2) 2
(J → J + 2) = Q
Pnl, n
l
(q) = | Ψnlm (r)|e iq·r/h̄
|Ψn
l
m
(r) |2
dk̂f 15 ki (2J + 1)(2J + 3)
m,m

which are all spherical symmetrical. The sum
with Ψ(r) = Rnl (r)Ylm (r̂) can be decomposed as
dσ (q) 4 2
(J → J, J ± 2) = Q
l+l

dk̂f 45
Pnl, n
l
(q) = (2l + 1)(2l
+ 1) (2L + 1)
L=|l−l
|
is independent of the initial value of J. The integral cross
 2 sections
L l l
(L)
2
   
× fnl, n
l
(q) 8π J +1 ki + kf
0 0 0 σ (d) (J → J + 1) = 2 ln D2
3ki 2J + 1 ki − kf
where {. . .} is the Wigner’s 3j-symbol and fnl,
L
n
l
(q) is
the radial integral 8π kf (J + 1)(J + 2) 2
σ (q) (J → J, J + 2) = Q

∞ 15 ki (2J + 1)(2J + 3)
(L)
fnl, n
l
(q) = Rnl (r)Rn
l
(r) jL (qr) r 2 dr all satisfy the Detailed Balance relation
0
ki2 (2Ji + 1)σ(Ji → Jf ) = kf2 (2Jf + 1)σ(Jf → Ji ) .
where jL is the modified Bessel function. For nlm →
n
l
m
subshell transitions, the amplitude decomposes as The summed diffusion cross sections are


dσ
l+l
σ (d) = [ (J → J
)](1 − cos θ) dk̂f ,
Fnlm, n
l
m
(q) = 4π
(L) (L)
ıL wlm
l
m
fnl, n
l
(q) YL,M (q̂) J±1 d k̂ f
 
L=|l−l
| 8π
= D2 ,
3ki2
where M = m − m
and where the coefficients
σ (q) = (16π/45)Q2 .
 1
(L) (2l + 1)(2l
+ 1)(2L + 1) 2
wlm
l
m
=
4π
   2.4.8 List of Born Cross sections for
L l l
L l l
Model Potentials
× .
M −m m
0 0 0
k2 = (2MAB /h̄2 )E, K = 2k sin 12 θ ,
Exact algebraic expressions for the probability
U = (2MAB /h̄2 )V , U/k2 = V /E .
Pn,n
(q) = | n
l
m
|eiq·r/h̄ |nlm |2 For a symmetric potential,
l
l
m,m


sin KR 2
fB (K) = − U (R) R dR .
of n → n
transitions in atomic hydrogen, have been KR
recently derived [11] as analytical functions of n and n
.
14 / Electron, Ion and Atom Scattering

Exponential: Polarization potential:

V (R) = V0 exp(−αR). V (R) = V0 /(R2 + R02 )2 .
2αU0  
fB (K) = − , 1 U0
(α2 + K 2 )2 fB (K) = − π exp(−KR0 ) ,
 4  4 R0
16 2 2 4  3  
2 3α + 12α k + 16k π U0 V0
σB (E) = πU0 σB (E) = [1 − (1 + 4kR0 ) exp(−4kR0 )] .
3 α4 (α2 + 4k2 )3 32R04 E
  
E→∞ 4 V0 U0
−→ π .
3 E α4

2.5 QUANTAL POTENTIAL

Gaussian: SCATTERING
V (R) = V0 exp(−α2 R2 )¿ The Schrödinger equation
 1/2   
π U0 h̄2
fB (K) = − exp(−K 2 /4α2 ) , − ∇ + V (r) Ψ+
2 +
4α2 k (r) = EΨk (r)
 2   2MAB r
π U0 V0
σB (E) = 1 − exp(−2k2 /α2 ) . solved subject to the asymptotic condition
8α4 E
1
k (r) ∼ exp(ik · r) + f (θ, φ) exp(ikr)
Ψ+
r
Spherical Well:
for outgoing spherical waves is equivalent to the solution
V (R) = V0 for R < a, V (R) = 0 for R > a. of the Lippman − Schwinger Integral Equation


U0
fB (K) = − [sin Ka − Ka cos Ka] , Ψ+k (r) = Φ +
k (r) + G(r, r
)U (r
)Ψ+

k (r ) dr


K3
π V0
σB (E) = (U0 a4 ) 1 − (ka)−2 + (ka)−3 sin 2ka where the outgoing Green’s Function for a free particle
2 E
is
− (ka)−4 sin2 2ka .
1 exp{ik|r − r
|}
G(r, r
) = − .
4π |r − r
|
Screened Coulomb Interaction:
Solution of the scattering amplitude may then be deter-
V (R) = V0 exp(−αR)/R. mined from the asymptotic form of Ψ+ k (r) directly or
from the integral representation
U0
fB (K) = − , 1
α2
+ K2 f (θ, φ) = − (2MAB /h̄2 ) exp(ikf · r)|V (r)|Ψ+
   4π i
4πU02 V0 U0
σB (E) = 2 2 →π .
α (α + 4k2 ) E α2 The differential cross section for elastic scattering is
where U0 = 2Z/a0 . As α → 0, then fB (θ) = −U0 /α . 2
dσ 2
= |f (θ, φ)| .
dΩ
Electron-Atom Model Static Interaction:
V (R) = −N (e2 /a0 ) [Z + a0 /R] exp(−2ZR/a0 ) . 2.5.1 Partial Wave Expansion
  A plane wave of unit amplitude can be decomposed
2N 2α2 + K 2
fB (θ) = , α = 2Z/a0 , according to
a0 (α2 + K 2 )2
πa20 N 2 12Z 4 + 18Z 2 k2 a20 + 7k4 a20 ∗
Φk (r) = eık·r = 4π ı" j" (kr)Y"m (k̂)Y"m (r̂)
σB (E) = .
3Z 2 (Z 2 + k2 a20 )3 ",m

The parameters for atomic H(1s) and He(1s2 ) are where j" is the spherical Bessel function which varies
(N = 1, Z = 1) and (N = 2; Z = 27/16), respectively. asymptotically as
 Encyclopedia of Chemical Physics and Physical Chemistry / 15

1 1 can be deduced. The Scattering, Transition and Re-

sin(kr − <π) .
j" (kr) ∼
kr 2 actance matrix elements are defined, in terms of the
The addition theorem for the spherical harmonics is phase shift η" suffered by each partial wave, as
∗
4π Y"m (k̂)Y"m (r̂) = (2< + 1)P" (k̂ · r̂) . S" (k) = exp(2iη" ) ,
m
T" (k) = 2i sin η" exp(iη" ) ,
Another useful identity is,
K" (k) = tan η" .
sin Kr ∗
= 4π j" (ki r)j" (kf r)Y"m (k̂i )Y"m (kˆf ) The asymptotic (kr → ∞) form of F" may then be
Kr
",m written in terms of the following linear combinations,
where K 2 = ki2 + kf2 − 2ki kf cos θ. The system wave-  
T"
F" (kr) ∼ sin(kr − <π/2) + ei(kr−"π/2) ,
function Ψ+k (r) ∼ N Φk with amplitude N is expanded 2i
according to
1
∗ = − [e−i(kr−"π/2) − S" ei(kr−"π/2) ] ,
Ψ+
k (r) = ı" eıη R#" (r)Y"m (k̂)Y"m (r̂) 2i
",m = eiη cos η" [sin(kr − <π/2) + K" cos(kr − <π/2)] .
4πN ∗
= ı" F#" (r)Y"m (k̂)Y"m (r̂) . expressed as a combinations of standing waves (trigono-
kr
",m metric functions), of incoming (-) and outgoing (+)
The radial wave F" is the solution of the radial spherical waves (exponential functions) and of a stand-
Schrödinger equation ing wave and an outgoing spherical wave. The physical
   significance of the admixture coefficients S" , T" and K"
d2 F" <(< + 1) is then transparent. The elements are connected by
+ k − U (r) +
2
F" (r) = 0 .
dr 2 r2
S" = 1 + T" = (1 + iK" )/(1 − iK" )
The reduced potential and energy are U (r) =
(2MAB /h̄2 )V (R) and k2 = (2MAB /h̄2 )E, respectively. K" is real while both S" and T" are complex. In term of
They both have dimensions of [a−2 0 ]. Also (ka0 )
2
= the full solutions F" of the radial Schrödinger equation,
(2E/ε0 )(MAB /me ). Each <-partial wave is separately the T-matrix element for elastic scattering is
scattered since the angular momentum of relative motion

is conserved for central forces. The radial waves R#" (r) 2i ∞ (0)
T" = − F" (r)U (r)F" (r) dr
and F#" (r) vary asymptotically as k 0
N" 1 (0)
R#" (r) ∼ sin(kr − <π + η" ) where F" = (kr)j" (kr) is the radial component of
r 2 the final plane wave. The Born approximation to T" is
1 (0)
F#" (r) ∼ eıη sin(kr − <π + η" ) . obtained upon the substitution F" = F" .
2
The amplitude of the partial radial wave wave R#" (r)
is N" = 4πN/k. In Table 5.1 are displayed the 2.5.3 Integral cross sections
amplitudes N and N" appropriate to various choices for ∞
4π
normalization of the continuum wavefunctions Ψk (r). σ(E) = (2< + 1) sin2 η" .
k2
"=0
∞
2.5.2 Scattering Amplitudes π
= 2 (2< + 1) |T" |2
k
"=0
For symmetric interactions V = V (r), the wavefunc- ∞
2π
tions Ψ+i and exp(ikf · r) are decomposed into partial = (2< + 1)[1 − ReS" ] .
waves. From their asymptotic forms, the following par- k
"=0
tial wave expansions for the scattering amplitude
∞
The semiclassical version is obtained by the substitution
1 mvb = (< + 12 )h̄ so that k2 b2 = (< + 12 )2 in terms of
f (θ) = (2< + 1) [exp(2iη" ) − 1] P" (cos θ) ,
2ik the impact parameter b. Regarding < as a continuous
"=0
∞ variable,
1
∞
∞
f (θ) = (2< + 1) [S" (k) − 1] P" (cos θ) , π
2ik σ(E) = 2 (2< + 1) |T" |2 d< = 2π |T (b)|2 b db .
"=0
∞
k 0 0
1
f (θ) = (2< + 1)T" (k)P" (cos θ) . The transition matrix |T (b)|2 is therefore the probability
2ik
"=0
of scattering of particles with impact parameter b.
16 / Electron, Ion and Atom Scattering

2.5.4 Differential cross sections × cos(η"+1 − η" )

∞
The differential cross section for elastic scattering is a2 = 5 b" sin2 η" + c" sin η" sin η"+2
"=0
dσ 2
= |f (θ)| = A(θ)2 + B(θ)2 × cos(η"+2 − η" )]
dΩ
where the real and imaginary parts of f (θ) are respec- where
tively
<(< + 1)(2< + 1)
∞ b" = ,
1 (2< + 1)(2< + 3)
A(θ) = (2< + 1) sin 2η" P" (cos θ)
2k 3(< + 1)(< + 2)
"=0 c" = .
∞ (2< + 3)
1
B(θ) = (2< + 1) [1 − cos 2η" ] P" (cos θ) .
2k
"=0

Their individual contributions to the integral cross sec- Example: S and P wave contributions
tions are

∞ The combined S, P wave (< = 0, 1) contributions to
4π
A(θ)2 dΩ = 2 (2< + 1) sin2 η" cos2 η" , the differential and integral cross sections are
k
"=0

∞ dσ 1
4π = 2 sin2 η0 + [6 sin η0 sin η1 cos(η1 − η0 )] cos θ
B(θ)2 dΩ = 2 (2< + 1) sin4 η" . dΩ k
k
"=0
+ 9 sin2 η1 cos2 θ ,
4π
σ(E) = 2 sin2 η0 + 3 sin2 η1 .
Expansion in Legendre Polynomials k
For pure S-wave scattering, the Differential cross
When expand as a series of Legendre Polynomials section (DCS) is isotropic. For pure P -wave scattering,
PL (cos θ), the differential cross section has the following the DCS is symmetric about θ = π/2, where it vanishes;
form the DCS rises to equal maxima at θ = 0, π. For combined
∞ S- and P -wave scattering, the DCS is asymmetric with
dσ(E, θ) 1
= 2 aL (E)PL (cos θ) forward-backward asymmetry.
dΩ k
L=0

where the coefficients

2.5.5 Optical Theorem
∞ "+L
aL = (2< + 1)(2<
+ 1)(<<
00 | <<
L0)2
The optical theorem relates the integral cross section
"=0 "
=|"−L|
to the imaginary part of the forward scattering amplitude
× sin η" sin η"
cos(η" − η"
) by
are determined by the phase shifts η" and the Clebsch-
σ(E) = (4π/k)Imf (0) .
Gordon Coefficients (<<
mm
| <<
LM ).
This relation is a direct consequence of the conservation
Example: Three-Term Expansion in cos θ of flux. The target casts a shadow in the forward direc-
tion where the intensity of the incident beam becomes re-
The differential cross section can be expanded as duced by just that amount which appears in the scattered
  wave. This decrease in intensity or shadow results from
dσ(E, θ) 1 1 3 interference between the incident wave and the scattered
= 2 (a0 − a2 ) + a1 cos θ + a2 cos2 θ .
dΩ k 2 2 wave in the forward direction. Figure 2.2 for the density
The coefficients are |Ψ+k (r)| of Section 2.5 illustrates how this interference
tends to illuminate the shadow region at the RHS-side of
∞
the target. Flux conservation also implies that the phase
a0 = (2< + 1) sin2 η" , shifts η" are always real. Thus
"=0
∞ 2 2
|S" | = 1, |T" | = ImT" .
a1 = 6 (< + 1) sin η" sin η"+1 ,
"=0
 Encyclopedia of Chemical Physics and Physical Chemistry / 17

(< + 3)(< + 4)
× sin2 (η" − η"+4 )
(2< + 5)

2(2<2 + 6< − 3)
+ sin (η" − η"+2 ) .
2
(2< − 1)

2.5.8 Born Phase Shifts

For a symmetric interaction, the Born amplitude is


sin KR 2
fB (K) = − U (R) R dR
KR
where U (r) = (2MAB /h̄2 ) = V (R). Comparison with
the partial wave expansion for fB (K) and
∞
sin KR
= (2< + 1) [j" (kR)]2 P" (cos θ)
KR
"=0
Figure 2.2. Scattering of an incident plane wave.
provides the Born phase shift

∞
2
2.5.6 Levinson’s Theorem tan η" (k) = −k
B
U (R) [j" (kR)] R2 dR .
0

For a local potential V (r) which supports n" bound

states of angular momentum < and energy En < 0, Examples of the Born S-wave Phase Shift
the phase shift limk→0 η" (k)) tends in the limit of zero
∞
collision energy to n" π. When the well becomes deep 1
tan η0B (k) =− U (R) sin2 (kR)dR
enough so as to introduce an additional bound level k 0
En+1 = 0 at zero energy, then limk→0 η0 (k) = (n0 + 12 )π.
e−αR
For the potential U = U0 ,
R
2.5.7 Partial Wave Expansion for U0
Transport cross sections tan η0B = − ln 1 + 4k2 /α2
4k
U0
The transport cross sections For the potential U = ,
(R2 + R02 )2
 −1
1 + (−1)n πU0
σ (n)
(E) = 2π 1 − tan η0B = − 1 − (1 + 2kR0 )e−2kR0 .
2(n + 1) 4kR03

+1
dσ
× [1 − cosn θ] d(cos θ) .
−1 dΩ
Born Phase Shifts (Large )
for n=1-4 have the following phase shift expansions
For <  ka,
∞
4π
∞
σ (1)
(E) = 2 (< + 1) sin (η" − η"+1 ) ,
2
k2"+1
k
"=0
tan η"B = − 2 U (R)R2"+2 dR ,
  ∞
[(2< + 1)!!] 0
4π 3 (< + 1)(< + 2)
σ (2)
(E) = 2 sin2 (η" − η"+2 ) , valid only for finite range interactions U (R > a) = 0.
k 2 (2< + 3)
"=0
∞ 
σ (3) (E) =
4π(< + 1) (< + 2)(< + 3)
sin2 (η" − η"+3 )
Example
k2
(2< + 5) (2< + 3)
"=0
 U = −U0 , R ≤ a and U = 0, R > a.
3(<2 + 2< − 1)
+ sin2 (η" − η"+1 ) , (ka)2"+1
(2< − 1) tan η"B (<  ka) = U0 a2 .
  ∞ [(2< + 1)!!] (2< + 3)
2
4π 5 (< + 1)(< + 2)
σ (4) (E) = 2
k 4 (2< + 3)(2< + 7)
"=0 The ratio η"+1 /η" ∼ (ka/2<)2 .
18 / Electron, Ion and Atom Scattering

2.5.9 Coulomb Scattering with even and odd values of the total spin St , respec-
tively. The scattering wavefunction for a pair of identical
For elastic scattering by the interaction V (r) = particles in spatially symmetric (+) or antisymmetric (-)
ZA ZB e2 /r, the Coulomb wave can be decomposed as states behaves asymptotically as

(C)
Ψk (r) =
4π ∗
ı" eıη F#" (r)Y"m (k̂)Y"m (r̂) ΨS,A (R) → [exp(ik · R) ± exp(−ik · R)]
kr exp(ikR)
",m
+ [f (θ, φ) ± f (π − θ, φ + π)]
R
where the radial wave varies asymptotically as
The differential cross section for scattering of both the
1 (C) projectile and target particles into direction θ is
F" ∼ sin(kR − <π + η" − β ln 2kR)
2  
dσ
where the parameter β is ZA ZB e2 /h̄v. The Coulomb = |f (θ, φ) ± f (π − θ, φ + π)|2
dΩ S,A
phase shift is
in the CM frame where scattering of the projectile into
(C)
η" = argΓ(< + 1 + iβ) = Im ln Γ(< + 1 + iβ) polar direction(π − θ, φ + π) is accompanied by scattering
of the identical target particle into direction θ, φ. This
to give the Coulomb S-matrix element is related to the probability that both identical particles
Γ(< + 1 + iβ) are scattered into θ. In the classical limit, where the
(C) (C)
S" = exp 2iη" = . particles are distinguishable, the classical cross section is
Γ(< + 1 − iβ)
 
dσ
The Coulomb scattering amplitude is = |f (θ, φ)|2 + |f (π − θ, φ + π)|2
dΩ C
(C)  
β exp 2iη" − iα ln sin2 12 θ the sum of the cross sections for observation of the
fC (θ) = − . projectile and target particles in the direction (θ, φ).
2k sin2 12 θ
Since P" [cos(π − θ)] = (−1)" P" (cos θ), the differential
2 cross section for φ-independent amplitudes f is then
The Coulomb differential cross section |fC | is
    ∞ 2
dσ Z 2 Z 2 e4 dσ 1  

= A B2 csc4 12 θ = 2 ω" (2<+1) [exp 2iη" −1] P" (cos θ) .
dΩ Coul 16E dΩ S,A 4k  
"=0

This is the Rutherford scattering cross section. It is For scattering in the symmetric (S) channel where St
interesting to note that Born and Classical theory also is even, ω" = 2 for < even and ω" = 0 for < odd. For
reproduce this cross section. Moreover scattering in the antisymmetric channel where St is odd,
  2 2 2
 2 2  ω" = 0 for < even and ω" = 2 for < odd. The integral
dσ 4MAB ZA ZB 2 2 ZA ZB
(q) = = 4a0 (MAB /me ) cross section is
dΩ Coul q4 (Ka0 )4
∞
8π
is a function only of the momentum transferred q = σS,A (E) = 2 ω" (2< + 1) sin2 η" .
k
h̄K = 2h̄k sin 12 θ in the collision. Note that q 2 = "=0
8MAB E sin2 12 θ. Let gA and gS be the fractions of states with odd and
even total spins St = 0, 1, 2, . . . , 2s. When the 2s+1 spin-
2.6 COLLISIONS BETWEEN states St are unresolved, the appropriate combination
IDENTICAL PARTICLES of symmetric and antisymmetric cross sections is the
weighted mean
The identical colliding particles, each with spin s,    
dσ dσ dσ
are in a resolved state with total spin St in the range = gS + gA
dΩ dΩ S dΩ A
(0 → 2s). The spatial wavefunction, with respect to par-
ticle interchange satisfies Ψ(R) = (−1)St Ψ(−R). Wave- σ(E) = gS σS (E) + gA σA (E)
functions for identical particles with even or odd total
spin St are therefore symmetric (S) or antisymmetric (A)
with respect to particle interchange. The appropriate 2.6.1 Fermion and Boson Scattering
combinations are ΨS,A (R) = Ψ(R) ± Ψ(−R), where the
positive sign (symmetric wavefunction S) and the nega- (a) For Fermions with half integral spin s, the
tive sign (antisymmetric wavefunction A) are associated statistical weights are gS = s/(2s + 1) and gA = (s +
 Encyclopedia of Chemical Physics and Physical Chemistry / 19

1)/(2s + 1). The differential cross section for fermion- (c) two spin-1 bosons (e.g. deuteron-deuteron)
fermion scattering is then
  dσ β2
dσF 2 = 2 csc4 12 θ + sec4 12 θ + 2
3 csc2 12 θ sec2 12 θ cos γ ,
2 2
= |f (θ)| +|f (π − θ)| − Re [f (θ)f ∗ (π − θ)] dΩ 4k
dΩ 2s + 1
where β = (Ze)2 /h̄v and γ = 2β ln(tan 12 θ). These are
The integral cross section fermion-fermion collisions is the Mott Formulae.
1 1
σF = [σS + σA ] − [σS − σA ] /(2s + 1) .
2 2 2.6.3 Scattering of Identical Atoms
which reduces, for fermions with spin 1/2, to
Two ground-state hydrogen atoms, for example, in-
 ∞
2π teract via the X 1 Σ+ 3 +
g and b Σu electronic states of H2 .
σF (E) = 2 (2< + 1) sin2 η" The nuclei are interchanged by rotating the atom pair
k
"=even
∞  by π, then by reflecting the electrons first through the
midpoint of R and then through a plane perpendicular
+3 (2< + 1) sin2 η" .
to the original axis of rotation. The mid-point reflection
"=odd
changes the sign only of the ungerade state wavefunction
(b) The statistical weights for bosons with integral and both Σ+ states are symmetric with respect to the
spin s, are gS = (s + 1)/(2s + 1) and gA = s/(2s + 1). plane reflection.
The differential cross section for boson-boson scattering The cross section for scattering by the gerade poten-
is tial is then the combination
       
dσB 2 dσ dσ dσ
Re [f (θ)f ∗ (π − θ)]
2 2
= |f (θ)| +|f (π − θ)| + . = 14 + 34
dΩ 2s + 1 dΩ g dΩ S dΩ A
The integral cross section boson-boson collisions is
of S and A cross sections which involve the phase shifts η"S
1 1 calculated under the singlet interaction. For scattering
σB = [σS + σA ] + [σS − σA ] /(2s + 1)
2 2 by the ungerade triplet interaction
which reduces, for bosons with zero spin, to      
dσ 1 dσ 3 dσ
 ∞  = 4 +
8π dΩ u dΩ A 4 dΩ S
σB (E) = 2 (2< + 1) sin2 η" .
k where the S and A cross sections involve the phase shifts
"=even
η"T calculated under the triplet interaction. Since the
Symmetry oscillations therefore appear in the differen- electrons have statistical weights 1/4 and 3/4 for the
tial cross sections for fermion-fermion and boson-boson 1 +
Σg and 3 Σ+ u states, the differential cross section for
scattering. They originate from the interference between H(1s) − H(1s) scattering by both potentials is
unscattered incident particles in the forward (θ = 0) di-
   
rection and backward scattered particles (θ = π, < = 0). dσ 1 dσ 3 dσ
A general differential cross section for scattering of spin = 4 + .
dΩ dΩ g 4 dΩ u
s particles is
The above combinations also hold for the integral cross
dσ (−1)2s
= |f (θ)|2 +|f (π − θ)|2 + 2Re [f (θ)f ∗ (π − θ)] . sections.
dΩ 2s + 1

Scattering of Incident Beam alone

2.6.2 Coulomb scattering of two
identical particles Since the current of incident particles ji = 2v, the
cross sections presented by the target (i.e., the number
(a) two spin-zero bosons (e.g. 4 He – 4 He) of incident particles removed from the beam in unit time
per unit incident current) are 12 of all those above. For
dσ β2 example,
= 2 csc4 12 θ + sec4 12 θ + 2 csc2 12 θ sec2 12 θ cos γ ,
dΩ 4k  I  
dσ 1 1 dσ
(b) two spin- 12 fermions (e.g. H+ – H+ , e± – e± )
2
= |f (θ) ± f (π − θ)| =
dΩ S,A 2 2 dΩ S,A
dσ β2
= 2 csc4 12 θ + sec4 12 θ − csc2 12 θ sec2 12 θ cos γ , I
and σS,A (E) = 12 σS,A (E).
dΩ 4k
20 / Electron, Ion and Atom Scattering

2.7 QUANTAL INELASTIC [∇2R + Kn2 (R)]Fn (R) = [Xnj · ∇R + Tnj (r)]Fj (R)
HEAVY-PARTICLE COLLISIONS j

The wavefunction for the complete A − B collision of coupled equations for the relative motion functions
system satisfies the Schrödinger equation Fn . The local momentum Kn is determined from Kn2 =
2MAB [E − En (R)]/h̄2 and the coupling matrix elements
h̄2 are
H(r, R)Ψ(r, R) = [Ĥint (r) − ∇2 + V (r, R)]Ψ(r, R)
2MAB R
Xnj (R) = −2 Φn (r, R)|∇R |Φj (r, R) r
= EΨ(r, R)
and
where the internal Hamiltonian is the sum Ĥint (r) =
HA (rA ) + HB (rB ) of individual Hamiltonians HA,B for Tnj (R) = − Φn (r, R)|∇2R |Φj (r, R) .
r
each isolated atomic or molecular species. The total
energy (internal plus relative) Solution of the above set for Fn (R) represents the
adiabatic close coupling method. The adiabatic states
h̄2 ki2 h̄2 kf2 are normally determined (via standard computational
E= + &i = + &f
2MAB 2MAB techniques of quantum chemistry) relative to a set of
axes (X
, Y
, Z
) with Z
-axis directed along nuclear
remains constant for all channels f throughout the col-
separation R. On transforming to this set which rotates
lision. The combined internal energy &j of A and B at
during the collision, then Ψ(r
, R
), for the diatomic
infinite separation R is &j (A)+&j (B) which are the eigen-
A − B case, satisfies
values of the internal Hamiltonian Ĥint corresponding to
the combined eigenstates ΦA (rA )ΦB (rB ). There are two h̄2
limiting formulations (diabatic and adiabatic) for describ- [Ĥ0 (r
) + V (r
, R
) − K̂]Ψ(r
, R
) = EΨ(r
, R
)
2MAB R
2
ing the relative motion. These depend on whether the
mutual electrostatic interaction V (r, R) between A and where the perturbation operator to the molecular wave-
B at nuclear separation R, or the variation in the kinetic functions in the rotating frame is
energy of relative motion, is considered to be a perturba-  
tion to the system i.e., on whether the incident speed vi ∂
2 ∂
K̂ = R − (L̂X
− JˆX
)2 − (L̂Y
− JˆY
)2
is fast or slow in comparison with the internal motions ∂R
∂R

e.g., with the electronic speed of the electrons bound to
A and B. in terms of the operators L̂ and Jˆ for the total and
internal angular momentum L and J respectively of
the collision system. Note LZ
= JZ
, for diatoms.
2.7.1 Adiabatic Formulation (Kinetic An advantage of using this rotating system in the
Coupling Scheme) adiabatic treatment is that radial perturbations, which
cause vibrational v → v
and electronic nl → n
l
When relaxation of the internal motion during the transitions, originate from the first term (radial) of
collision is fast compared with the slow collision speed K̂ while angular perturbations (torques) which causes
vi , or when the relaxation time is short compared rotational J → J
and electronic nl → nl
transitions
with the collision time , the kinetic energy operator originate form the angular momentum operator products
(2MAB /h̄2 )∇2R is then considered as a small perturba- [L̂X
JˆX
+ L̂Y
JˆX
]. The use of a rotating frame causes
tion to the quasi-molecular A − B system
 at fixed R. some complication, however, to the direct use of the
The system wavefunction Ψ(r, R) = n Fn (R)Φn (r, R) asymptotic boundary condition for Ψ(r
, R
).
is therefore be expanded in terms of the known “adia-
batic” molecular wavefunctions Φn (r, R) for the quasi-
molecule AB at fixed nuclear separation R. This set of 2.7.2 Diabatic Formulation (Potential
orthornormal eigenfunctions satisfies Coupling Scheme)
[Ĥint (r) + V (r, R)]Φn (r, R) = En (R)Φn (r, R) . When relaxation of the internal motion is slow com-
pared with the fast relative speed vi , then Ψ is expanded
As R → ∞, both Φn (r, R) and the eigenenergies En (R) in terms of the known unperturbed (diabatic) orthonor-
tend in the limit of infinite nuclear separation R to the mal eigenstates Φj (rA , rB ) = ψi (rA )φk (rB ) of Ĥint ac-
(diabatic) eigenfunctions Φn (rA , rB ) = ψi (rA )φj (rB ), of cording to
Ĥint with eigenenergies
 &n , respectively. The substitu-
tion Ψ(r, R) = n Fn (R)Φn (r, R) into the Schrödinger Ψ(r, R) = Fj (R)Φj (r) .
equation results in the following set j
 Encyclopedia of Chemical Physics and Physical Chemistry / 21

Substituting into the Schrödinger equation, multiplica- 2.7.3 Inelastic scattering by a Central
tion by Φ∗n (r) and integration over r, the unknown func- Field
tions Fn (R) for the relative motion in channel n are then
found to satisfy the infinite set of coupled equations When the atom-atom or atom-molecule interaction
is spherically symmetric in the channel vector R i.e.,
[∇2R + Kn2 (R)]Fn (R) = Unj (R)Fj (R) . V (r, R) = V (r, R), then the orbital l and rotational
j=i j angular momenta are each conserved throughout the
collision so that a <-partial wave decomposition of the
The reduced potential matrix elements which couple the
translational wavefunctions for each value of j is possible.
internal states n and j are
The translational wave is decomposed according to
2MAB ∗ 4πN
Unj (R) = Vnj (R) = Ujn (R) ∗
h̄2 Fj (R) = ı" Fj" (R)Y"m (k̂)Y"m (R̂)
ki R
",m
where the electrostatic interaction averaged over states n
and j is and inserted into the diabatic set of coupled equations (of
Section 2.7.2). The radial wavefunction Fj" is then the


solution of
Vnj (R) = Φ∗n (r)V (r, R)Φj (r) dr .  
d2 Fj" <(< + 1)
+ ki − {Uii (R) +
2
} Fj" (R)
The local wavenumber Kn of relative motion under the dR2 R2
static interaction Vnn is given by = Uij (R)Fj" (R)
j=i
Kn2 (R) = kn2 − Unn (R) .
which is the direct generalization of the quantal radial
The diagonal elements Unn are the distortion matrix ele- equation for potential scattering to directly include other
ments which distort the relative motion from plane waves channels j = i. The coupled equations are now solved
in elastic scattering, while the off-diagonal matrix ele- subject to the requirements that
ments, Uif and Uij , Ujf which couple states i and f  "
Tii
either directly or via intermediate channels j cause in- Fi" (ki R) ∼ sin(ki R − <π/2) + ei(ki R−"π/2)
elastic scattering and polarizations contributions to elas- 2i
tic scattering. In contrast to the adiabatic formulation, for the elastic scattered wave and
radial and angular transitions originate in the diabatic  
ki 1 Tij"
formulation from the radial and angular components to Fj" (kj R) ∼ ( ) 2 ei(kj R−"π/2)
the potential coupling elements Vnj (R). The set of cou- kj 2i
pled are solved subject to the usual asymptotic (R → ∞)
for the inelastic wave. The transition-matrix elements
requirement that
for elastic and inelastic scattering are

∞
Fj (R) ∼ exp(iki Z)δij + fij exp(ikj R)/R 2i (0)
Tij" = − 1 Fj" (r)Uji (r)Fi" (r) dr
for the elastic i = j and inelastic i = j scattered waves. (ki kj ) 2 0
In terms of the amplitude fij for scattering into direction (0)
(θ, φ), the differential and integral cross sections for i → j where Fj" = kf rj" (kf r) and Fi" (r) are the solutions of
transitions are the above coupled radial equations. The differential cross
section for inelastic scattering is
dσij vj
|fij (θ, φ)|
2 ∞ 2
=  
dΩ vi dσij  
= (1/4ki2 )  (2< + 1)Tij" P" (cos θ) .
dΩ  
and "=0

π
2π The integral inelastic cross section is

vj 2
σij = d(cos θ) |fij (θ, φ)| dφ . ∞
vi π  2
(2< + 1) Tij"  .
0 0
σij (E) =
ki2
As well as obtaining the scattering amplitude from the "=0
above asymptotic boundary conditions, fif can also
The transition matrix T" = {Tij" } is symmetrical, Tij" =
be obtained from the integral representation for the "
Tji , and the cross sections satisfy detailed balance. Each
scattering amplitude is  2
transition matrix element Tij"  is the probability of an
fif (θ) = Φf (r) exp(ikf · R)|V (r, R)|Ψ(r, R) r,R . i → f transition in the target for each value < of the
(orbital) angular momentum of relative motion.
22 / Electron, Ion and Atom Scattering

2.7.4 Two State Treatment Uii = Uf f = U , Uif = Uf i . Then the equations

can be decoupled by introducing the linear combinations
Here all couplings are ignored except the direct ψ ± (R) = √12 [ψi (R) ± ψf (R)], so the two-state set can
couplings between the initial and final states as in a two- be converted to two one-channel decoupled equations
level atom. The coupled equations to be solved are
[∇2 + k2 − (U ± Uif )]ψ ± (R) = 0 .
[∇ + 2
ki2 − Uii (R)]ψi (R) = Uif (R)ψf (R)
The problem has therefore been reduced to potential
scattering by the interactions U± = (U ± Uif ) associated
2
[∇ + kf2 − Uf f (R)]ψf (R) = Uf i (R)ψi (R) .
with elastic scattering amplitudes f ± . Hence the elastic
(i = f ) and ”inelastic” (i = f ) amplitudes are

Distorted-Wave Approximation fii = (f + + f − )/2 fif = (f + − f − )/2 .

Here all matrix elements in the two level equations In terms of the phase shifts ηl± associated with potential
(section 2.7.4) are included, except the back coupling scattering by U± , the amplitudes for elastic and inelastic
Vif Ψf term which provides the influence of the inelastic scattering are then
channel on the elastic channel and is required to conserve ∞
1 + −
probability. Distortion of the elastic and outgoing fin (θ) = (2l + 1)[(e2iηl + e2iηl )/2 − 1]Pl (cos θ)
inelastic waves by the averaged (static) interactions Vii 2ik
l=0
and Vf f respectively is therefore included. The two state
equations can then be decoupled and effectively reduced and
to one-channel problems. An analogous static-exchange ∞
1 + −
distortion approximation, where exchange between the fif (θ) = (2l + 1)[(e2iηl − e2iηl )/2 − 1]Pl (cos θ) .
incident and one of the target particles also follows from 2ik
l=0
the two-level treatment.
The corresponding differential cross sections |fif |2 will
therefore exhibit interference oscillations. The integral
Born Approximation cross sections are
∞
Here the distortion (diagonal) and back coupling 4π
σii = (2l+1)[(sin2 ηl+ +sin2 ηl− )/2−sin2 (ηl+ −ηl− )/4]
matrix elements in the two level equations (section 2.7.4) k2
l=0
are ignored so that ψi (R) = exp(iki · R) remains an
undistorted plane wave. The asymptotic solution for ψf and
when compared with the asymptotic boundary condition ∞
π
then provides the Born elastic (i = f ) or inelastic σif = (2l + 1) sin2 (ηl+ − ηl− ) ,
scattering amplitudes k2
l=0


1 2MAB respectively.
B
fif (θ, φ) = − Vf i (R)eik·R dR .
4π h̄2
The momentum change resulting from the collision in Examples: Atomic Collisions with Identical
Q = h̄K where k = ki − kf . The Born amplitude Nuclei
also follows by inserting Ψ(r, R) = Φi (r) exp i(ki · R) in
the integral representation. Comparison with potential Important cases of exact resonance are the Symmet-
scattering shows that the elastic scattering of structured rical Resonance Charge Transfer collision
particles occurs in the Born approximation via the
f (1s) + Hes (1s ) → Hef (1s ) + Hes (1s)
He+
averaged electrostatic interaction Vii (r). 2 2 +

For electron-ion or ion-ion collisions, the plane waves

exp(iki,f · R) are simply replaced by Coulomb waves to which converts a fast ion beam f to a fast neutral beam
provide the Coulomb-Born Approximation. and the excitation transfer collision

He(1s2s3 S) + He(1s2 1 S) → He(1s2 1 S) + He(1s2s3 S)

2.7.5 Exact Resonance
which transfers the internal excitation in the projectile
The two-state equations of section 2.7.4 cannot in beam fully to the target atom. The electronic molecular
general be solved analytically except for the specific wavefunctions divide into even (gerade) or odd (unger-
case of Exact Resonance when ki = kf = k and ade) classes upon reflection about the mid-point of the
 Encyclopedia of Chemical Physics and Physical Chemistry / 23

internuclear line (R → −R). In the separated atom atom-molecule scattering to a corresponding working set
limit,Ψg,u ∼ φ(rA ) ± φ(rB ).The potentials U± in the for- of coupled radial equations, analogous to those in 2.7.3,
mer case arethe gerade and ungerade interactions Vg,u . the orbital angular momentum l of relative motion must
The phase shifts for elastic scattering by the resulting be distinguished from the combined internal angular mo-
gerade (g) and ungerade (u) molecular potentials of A+ 2 mentum j associated with the internal (rotational and
are, respectively, η"g and η"u . The charge transfer (X) and electronic) degrees of freedom of the partners A and B
transport cross sections are then at rest at infinite separation R. Both the orbital angular
∞ momentum l of relative motion and the internal angular
π
σX (E) = (2< + 1) sin2 (η"g − η"u ) , momentum j of the atomic electrons or of molecular rota-
k2 tion are in general coupled. The total angular momentum
"=0
∞ J = l+j and its component Jz along some fixed direction
4π
(1)
σu,g (E) = (< + 1) sin2 (β" − β"+1 ) , (of incidence) are each conserved. Angular momentum
k2
"=0 may therefore be exchanged between the internal (rota-
  ∞ tional) and translational (orbital) degrees of freedom via
4π 3 (< + 1)(< + 2)
(2)
σu,g (E) = 2 sin2 (β" − β"+2 ) . the couplings Vnm (R) or K̂. Partial wave analysis is an
k 2 (2< + 3)
"=0 exercise in angular momentum coupling and is well estab-
lished (e.g., Ref.[14]) for both the diabatic and adiabatic
For ungerade potentials, β" = η"g for < even and η"u tratments of heavy-particle collisions.
for < odd. For gerade potentials, β" = η"u for < even and
(1)
η"g for < odd. The diffusion cross section σu,g contains
(2) 2.8 ELECTRON-ATOM INELASTIC
(g/u) interference. The viscosity cross section σu,g does
not. For charge transfer between the heavier rare gas ions COLLISIONS
Rg + with their parent atoms Rg, the degenerate states
at large internuclear separations are not s states but p 2.8.1 Close Coupling Equations for
states. The states are then Σg,u which arise from the electron-atom (ion) collisions
p state with m = 0 and Πg,u which arises from m =
±1 with space quantization along the molecular axis. A partial wave decomposition provides the full close-
Since there is no coupling between molecular states of coupling quantal method for treating A − B collisions,
different electronic angular momentum, the scattering by electron-atom, electron-ion or atom-molecule collisions.
the 2 Σg,u pair and the 2 Πg,u pair of N e+2 potentials (for
The method[15] is summarized here for the inelastic
example) is independent. The cross section is therefore processes
the combination
e− + Ai → e− + Af
1 2
σel,X (E) = σΣ (E) + σΠ (E)
3 3 at collision speeds less or comparable with those target
of cross sections σΣ and σΠ for the individual contribu- electrons actively involved in the transition. It is based
tions arising from the isolated 2 Σg,u and 2 Πg,u states to upon an expansion of the total wavefunction Ψ for the
elastic el or charge-transfer X scattering. See Refs.[12, 13] (e− − A) - multielectron system in terms of a sum of
for further details on excitation-transfer and charge- products of the known atomic target state wavefunctions
transfer collisions. |Φi > and the unknown functions Fi (r) for the relative
motion. Here


Singlet-Triplet spin flip cross section 1
Ψ (r1 , . . . , rN ; r) = A
Γ
ΦΓi (r1 , r2 , . . . rN ; r̂) FiΓ (r)
i
ki
This cross section is
∞ involves a sum over all discrete and an integral over
π
σST (E) = (2< + 1) sin2 (η"s − η"t ) the continuum states of the target. The operator A
k2 antisymmetrizes the summation with respect to exchange
"=0
of all pairs of electrons in accordance with the Pauli
where η"s,t
are the phase shifts for individual potential
exclusion principle. The angular and spin momenta
scattering by the singlet and triplet potentials, respec-
(denoted collectively by r̂) of the projectile electron have
tively.
been coupled with the orbital and spin angular momenta
of the target states |Φi > to produce the ”channel
2.7.6 Partial Wave Analysis functions” ΦLSπ
i (r1 , r2 , . . . , ri ; r̂) which are eigenstates
of the total orbital L, total spin S angular momentum,
In order to reduce the three dimensional diabatic or their Z - components ML , MS and parity π. The set
adiabatic set of coupled equations for atom-atom and Γ ≡ LSML MS π of quantum numbers are therefore
24 / Electron, Ion and Atom Scattering

conserved throughout the collision. By substituting the The integral cross section for the transition i ≡ αa Li Si →
expansion for ΨΓ into the Schrödinger Equation, f ≡ αf Lf Sf in the target atom, where α denotes
  the additional quantum numbers required to completely
N +1   N +1
1 Z 1  Γ specify the state, is then
HN +1 ΨΓ =  − ∇2i − + Ψ
2 r i r
i>j=1 ij π (2L + 1)(2S + 1)
i=1
σif (ki2 ) = |T Γ |2 .
ki2 2(2Li + 1)(2Si + 1) ij
LSπ
expressed in atomic units, the radial functions for the
motion of the scattered electron satisfy the infinite set of According to detailed balance, the collision strength
coupled integro-differential equations
Ωif = ki2 (2Li + 1)(2Si + 1)σif (ki2 )
 2 
d <i (<i + 1) 2(Z − N )
+ ki −
2
+ FiΓ (r) is therefore dimensionless and is symmetric with respect
dr 2 r2 r to i → f interchange. Further extensions, simplifications
=2 VijΓ (r) + WijΓ (r) FjΓ (r) . and calculational schemes of the basic close coupling and
j related methods are found in Refs.[15–17].
With modern high speed computers, it is feasible
The direct potential couplings are represented by to solve the coupled set of radial equations only for a
 N
 restricted basis set of unperturbed states Φn (r) regarded
Zt 1
Γ
Vij (r) = Zp δij + Φi |
Γ
|Φ Γ
. as being closely and strongly coupled. For electron-
r |rk − r| j atom (molecule) collisions at low energies E, the full
k=1
quantal close coupling method is extremely successful in
The non-local exchange couplings are represented by
predicting the cross sections and shapes and widths of
N resonances which appear at energies E just below the
1
Wij FjΓ (r) = ΦΓi | |(A − 1)ΦΓj FjΓ . various thresholds for excitation of the various excited
|rk − r| levels. As E increases past the threshold for ionization, it
k=1
becomes less successful, and is plagued by problems with
The direct potential gives rise to the long-range polar-
convergence both in the number of the basis states and in
ization attraction which is very important for low energy
the number of partial waves used in the expansion for ΨΓ .
scattering. The exchange potentials are short range and
Other methods for the intermediate and high energies are
are extremely complicated. Additional non-local poten-
therefore preferable. For heavy particle collisions and for
tials that arise from various correlations (which cannot
electron collisions at high energies, semiclassical versions
be included directly but which can be constructed from
(in section (2.9)) of the close coupling equations can be
pseudostates) can also be added to the RHS of the equa-
derived.
tions.
Numerical solution of the above set of close coupled
equations is feasible only for a limited number of close 2.8.2 Close Coupling with Pseudostates
target states. For each N, several sets of independent and Correlation
solutions Fij of the resulting close coupled equations are
determined subject to Fij = 0 at r = 0 and to the Pseudostates: A partial acknowledgment of the in-
reactance K-matrix asymptotic boundary conditions, fluence of higher discrete and continuum states, not in-
cluded within the wave function expansion, is to add, to
FijΓ ∼ sinθi δij + Kij
Γ
cosθi the truncated set of basis states, functions of the form
for n open channels characterized by ki2 = 2(E − Ei ) > 0. ΨP (r)ΦP (r) where ΦP is not an eigenfunction of the in-
The argument is ternal Hamiltonian Ĥint but is chosen so as to repre-
sent some appropriate average of bound and continuum
1 Z −N states. These pseudostates can provide full polarization
θi = ki r − <i π + ln(2ki r) + σi
2 ki distortion to the target by incident electrons and allows
flux to be transferred from the the open channels included
where <i is the orbital angular momentum of the scat-
in the truncated set.
tered elecron and where σi = argΓ[<i + 1 − i(Z − N )ki ]
Correlation: When the initial and final internal states
is the Couloumb phase. For closed channels, ki2 < 0 and
of the system are not well separated in energy from other
Fij ∼ Cij exp(−|ki |r) as r → ∞. The scattering ampli-
states then the closed coupling calculation converges
tude can then be expressed in terms of the elements Tij
very slowly. An effective strategy is to add a series of
of the (n × n) T-matrix which is related to the K and S
correlation terms involving powers of the distance rij
matrices by,
between internal particles of projectile and target to
2ikΓ the truncated close coupling expansion which already
TΓ = = SΓ − I .
I − iKΓ includes the important states.
 Encyclopedia of Chemical Physics and Physical Chemistry / 25

2.8.3 The R-Matrix Method the full quantal expression for the cross section.
(C) The use of JWKB-approximate solutions of the ra-
This method, introduced originally in an analysis of dial Schrödinger Equation for the radial wavefunction
nuclear resonance reactions, has been extensively devel- R#," for A − B relative motion within the full quantum
oped [15–17] over the past twenty years as a powerful ab treatment of the A − B collision.
initio calculational tool. It partitions configuration space
into two regions by a sphere of radius r = a, where r is
the scattered electron coordinate. In the internal region 2.9.1 Classical Path Theory
r < a, the electron-atom complex behaves almost as as
The basic assumption here is the existence over the
a bound state so that a configuration interaction expan-
inelastic scattering region of a common classical trajec-
sion of the total wave function Ψ, as in atomic structure
tory R(t) for the relative motion under an appropri-
calculations, is appropriate. In the external region the
ately averaged central potential V̄ [R(t)]. The interac-
scattered electron moves in the long range multipole po-
tion V [r, R(t)] between A and B may then be considered
tential contained in the direct electrostatic interaction,
as time-dependent. The system wavefunction therefore
and can be accurately represented by a perturbation ap-
satisfies
proach. See Refs.[15–17] for further details, for other
modern quantal approximations and for various compu- ∂Ψ(r, t)
tational methods useful for electron-atom collisions over ih̄ = [Ĥint (r) + V (r, R(t))]Ψ(r, t)
∂t
a wide energy range.
and can be expanded in terms of the eigenfunctions Φn
of Ĥint as
2.8.4 Electron-Molecule Collisions
Ψ(r, t) = An (t)Φn (r) exp(−iEn t) .
The close coupling equations are also applicable to n
electron molecule collision but severe computational dif-
The transition amplitudes An then satisfy the set
ficulties arise due to the large number of rotational and
vibrational channels that must be retained in the ex- ∂An (b, t)
pansion for the system wavefunction. In the fixed nu- ih̄ = Aj (b, t)Vnj (R(t) exp(iωnj t)
∂t n
clei approximation, the Born-Oppenheimer separation
of electronic and nuclear motion permits electronic mo- of first order equations coupled by the matrix elements
tion and scattering amplitudes fnn
(R) to be determined Vnj (R) between states n and j with energy separation
at fixed internuclear separations R. Then in the adia- h̄ωnj = En − Ej . Once the classical trajectory R ≡
batic nuclear approximation the scattering amplitude for (R(t), θ(t), φ = const.) is determined from the classical
i ≡ n, v, J → n
, v
, J
≡ f transitions is: equations
fif () = χn
v
(R)YJ
MJ
(R̂)|fnn
(r)|χnv (R)YJMJ (R̂) dR 1
= ±v[1 − b2 /R2 − V̄ (R)/E] 2
dt
and cross sections can be obtained. See Ref.[15] for db vb
further details. = 2
dt R
of motion for impact parameter b and kinetic energy E =
2.9 SEMICLASSICAL INELASTIC 1 2
2 MAB v , the coupled equations are solved subject to the
SCATTERING requirement An (b, t → −∞) = δni . Since the probability
2
for an i → f transition is Pif = |Af (b, t → ∞| , the
The term semiclassical is used in scattering theory to differential cross section for inelastic scattering is
denote many different situations.  
(A) The use of some time-dependent classical path R(t) dσif dσel
= Pif (bn , φ)
within a time dependent quantal treatment of the re- dΩ dΩ
n
sponse of the internal degrees of freedom of A and B to
the time-varying field V (R(t) created by the approach where dσel /dΩ is the differential cross section
of A towards B along the classical trajectory R(t). This |bdb/d(cos θ)| for elastic scattering by V̄ (R) and where
procedure generalizes classical theory for potential scat- the summation is over all trajectories bn which pass
tering to structured collision partners and inelastic tran- through (θ, φ). The integral cross section is
sitions.
∞
(B) The use of the three-dimensional Eikonal − 2
σif = 2π |Af (b, ∞)| bdb .
Phase S (R), which is the solution of the Hamilton-Jacobi 0
Equation, for the channel wavefunction Ψ+ (R), within
26 / Electron, Ion and Atom Scattering

Impact Parameter Method 2.9.3 Eikonal Theories

Here the relative motion wavefunction Fn (R) is de-
This normally refers to the use of the straight line
composed as [22]
trajectory R(t) = (b2 + v 2 t2 )1/2 , θ(t) = arctan(b/vt)
within the classical path treatment. See Bates [18, 19] Fn (R) = An (R) exp iSn (R) exp(−χn (R)) .
for examples and further discussion.
The classical action, or solution of the Hamilton-Jacobi
Equation ∇Sn (R) = κn (R), for relative motion under
2.9.2 Landau - Zener Cross Section the channel interaction Vnn (R), is

R
The Landau - Zener transition probability is derived Sn (R) = k n · R + (κn − kn ) · dRn
R0
from an approximation to the full two - state impact-
parameter treatment of the collision. The single passage where R0 is the initial point on the associated trajectory
probability for a transition between the diabatic surfaces Rn (t) where κn = kn . The current Jn in channel
H11 (R) and H22 (R) which cross at RX is the Landau - n, assumed elastic, satisfies the conservation condition
Zener transition probability ∇Jn = 0, so that χn is the solution of

P12 (RX ; b) = 1 − exp(−2π|H12 (RX )|2 /h̄vX |H11

− H22 |) ∇2R Sn − 2(∇R Sn ) · (∇R χn ) = 0 .

Flux in channel n is therefore lost only via transition

where H12 is the interaction coupling states 1 and 2. The to another state f with probability controlled solely by
diabatic curves are assumed to have linear shapes in the Af . When many wavelengths of relative motion can be



vicinity of the crossing at RX i.e.,(H11 − H22 = ∆F ) and accomodated within the range of Vnn ,as at the higher
H12 is assumed constant. The adiabatic surfaces energies favored by the Diabatic Scheme, the fast R-
variation of Fn is mainly controlled by Sn , and the
1 1
W± = (H11 + H22 ) ± [(H11 − H22 )2 + 4H12 ] original Diabatic set of coupled equations then reduce
2 2 to the simpler set
do not cross (avoided crossing). They are separated at ∂Af (t)
RX by W + −W − = 2H12 . The probability for remaining ih̄ = An (t) V f n (Rf (t) exp i(Sn − Sf )
∂t
on the adiabatic surface is P12 (RX ). The probability for n=f

remaining on the diabatic surface or for pseudocrossing × exp −(χn − χf )

between the adiabatic curves is 1 − P12 (RX ). The overall
transition probability for both the incoming and outgoing of first-order coupled equations.When a common trajec-
legs of the trajectory R(t) is then tory Rn (t) = R(t) under some averaged interaction V̄ (R)
can be assumed for all channels n then

t
P12 = 2P12 (1 − P12 ) . −1
Sn (R)−Sf (R) = ωf n t+h̄ [Vf f (R(t))−Vnn (R(t))] dt
t0
The Landau-Zener cross section is
and the classical path equations are recovered.

RX
σ12 = 4π P12 (1 − P12 ) bdb
Averaged Potential
0
The orbit common to all channels is found by choos-
where the variation of P12 on impact parameter b arises ing the potential governing the relative motion as the
from the speed vX ≈ v0 (1 − b2 /RX 2 1/2
) at the crossing average[22]
point RX . For rectilinear trajectories R = b + v0 t ,
V(R) = Ψ(r, R)|Ĥint (r) + V (r, R)|Ψ(r, R) r .
2
σ12 (v0 ) = 4πRX [E3 (α) − E3 (2α)]
Hamilton’s equations of motion for this interaction
∞ −n  
where En (α) = 1 y exp(−αy) dy is the exponential
integral with argument α = 2π|H12 |2 /h̄v0 ∆F . See V(R) = |An |2 + A∗f An Vf n exp i(Sn − Sf )
Nikitin [20, 21] for more elaborate models which include n f
interference effects arising from the phases or eikonals
associated with the incoming and outgoing legs of the are therefore coupled to the set of first order equations for
trajectory. the transition amplitudes Af (R). An essential feature is
 Encyclopedia of Chemical Physics and Physical Chemistry / 27

that total energy is always conserved, being continually 0

redistributed between the relative motion and internal
10 6-State MSO/PRA 10 00
Q10t00_22state
degrees of freedom, as motion along the trajectory -2 Q10t00_22state
proceeds. In terms of the solutions Af (bj , t) and the 10 CC (Balakrishnan, Forrey, Dalgarno)
(j)
differential cross section dσel /dΩ for elastic scattering -4
of particles with impact parameter bj (θ) through θ by 10

)
V̄ (R), the semiclassical scattering amplitude is

2
-6

00 (Å
j


(j)
1/2 10
fif (θ) = Af [bj (θ), ∞] exp iSf (bj (θ), ∞) dσel /dΩ .
-8
10

Q10
The accumulated classical action for orbit bj (θ) is

R(t) -10


Sf (bj , t) = [κf (R) − kn (R)] · dR . 10
R0
-12
When the same scattering angle originates from more 10
than one impact parameter bj , then interference effects
-14
originate from the different actions associated with the 10
different orbits bj (θ). The contributions arising from N- -8 -6 -4 -2 0 2
orbits which are well separated combine according to 10 10 10 10 10 10
N E (eV)
j
fif (θ) = −i αj βj fif (θ) .
j=1

The coefficients αj = exp ±π/4 depends on whether the Figure 2.3. Vibrational Relaxation Cross Sections
scattered particle emerges on the same side (+) of the (Quantal and Semiclassical) as a function of Collision
axis as it entered, as in a collision overall repulsive, or on energy E.
the opposite side (-), as in an overall attractive collision.
The coefficients βj is exp ±π/4 according to whether the Multichannel Eikonal Method
sign of db/dθ is (+) or (-). The differential cross section
will therefore exhibit characteristic oscillations, directly For electronic transitions in electron-atom and heavy
attributable to interference between the action phases particle collisions at high impact energies, the major con-


Sf (bj ) associated with each contributing classical path tribution to inelastic cross sections arises from scattering
bj (θ). Th analysis can be extended, as in the uniform in the forward direction.The trajectories implicit in the
Airy function approximation to cover orbits which are not action phases and set of coupled equations can be taken
widely separated, as for the case of rainbow scattering or as rectilinear. The integral representation
of caustics, in general, where the density of paths become
fif (θ) = φf (r) exp(ikf · R)|V (r, R)|Ψ(r, R) r,R
infinite. The theory above provides the basis of the mul-
tistate orbital treatment[22] successful for rotational and for the scattering amplitude, where
vibrational excitation in atom-molecule and ion-molecule
collisions at higher energies EAB ≥ 11eV. Other semi- Ψ(r, R) = An (R)φn (r) exp iSn (R)
classical treatments based on the JWKB-approximation n
to the corresponding set of coupled equations for the ra- then provides the basis of the multichannel eikonal
dial wavefunction for relative motion can be found in treatment[28] valuable, in particular, for heavy particle
Refs.[23–25]. collisions and for electron (ion)-excited atom collisions
In Fig. [2.3] are displayed cross sections for the where, due to the large effect of atomic polarization
quenching process, (charge-induced dipole), the collision is dominated by
He + H2 (v = 1, J = 0) → He + H2 (v = 0, J = 0) scattering in the forward direction.

for collision energies E ranging from the ultracold to

1Kev. The full quantal results [26] are shown together 2.10 LONG RANGE INTERACTIONS
with those calculated [27]from the semiclassical Mulistate
Orbital Method [22]. It is seen that results from both 2.10.1 Polarization, Electrostatic and
methods complement and connect with each other very Dispersion Interactions
well, in that the quantal treatment is calculationally
feasible up to E − 1Ev while semiclassical procedures The long-range interaction V (R) between two
are feasible for the higher collision energies. atomic/molecular species can be decomposed into
28 / Electron, Ion and Atom Scattering

Vpolarization (R) + Velectrostatic (R) + Vdispersion (R) C6 C8 C10

Vdispersion ∼ − − 8 − 10 . . .
R6 R R
The Polarization Interaction arises from the interac-
tion between the ion of charge Ze and the multipole mo- represents the interaction of the fluctuating dipole inter-
ments it induces in the atom or molecule AB. The dom- acting with the induced dipole C6 -term and quadrupole
inant polarization interaction is the ion - induced dipole C8 -term, respectively. The leading R−6 -term represents
interaction the Van-der Waal’s attraction.

αd (Ze)2
Vpol (Ze; ind D) = − [1 + (αd
/αd )P2 (ŝ · R̂)]
2R4 REFERENCES
where the averaged dipole polarizability is αd = (α +
α⊥ )/3 and α and α⊥ are the polarizabilities of AB in 1. G. L. Cathen, J. Husain and R. N. Zare, J. Chem.
the directions parallel and perpendicular to the molec- Phys. 69 1737 (1978).
ular axis ŝ. The anisotropic polarizability is αd
= 2. E.A.Mason and E.W.McDaniel, Transport Properties
2(α − α⊥ )/3. The next polarization interaction is the of Ions in Gases (Wiley, New York, 1988).
charge - induced quadrupole interaction, averaged over 3. L.A. Viehland, E.A. Mason, W.F. Morrison and
all molecular orientations M.R. Flannery, At. Data Nucl. Data Tables 16 495
(1975).
Vpol (Ze; ind Q) = −ᾱq (Ze)2 /2R6 4. Introduction to the Quantum Theory of Scattering,
L. S. Rodberg and R. M. Thaler, (Academic Press,
where ᾱq is the averaged quadrupole polarizability. Addi- New York, 1967), p 226.
tional polarization terms arise from permanent multipole 5. The Theory of Atomic Collisions, N. F. Mott and H.
moments of one partner and the dipole (or multipole) it S. W. Massey, (Clarendon Press, Oxford, 1965), 3rd
induces in the other, averaged over all directions ed.
6. Collision Theory, M.L. Goldberger and K. M. Wat-
1
Vpol (D; ind D) = − (D2 ᾱdi + Di2 ᾱdn ) son, (Wiley, New York, 1964).
R6 n 7. Scattering Theory of Waves and Particles, R. G.
The Electrostatic Interaction results from the inter- Newton, (Mc Graw-Hill, New York, 1966).
action of the ion with the permanent multipole moments 8. Quantum Collision Theory, C. J. Joachain, (North-
of the neutral. For cylindrically symmetric neutrals or Holland, Amsterdam, 1975) p 383.
linear molecules, the ion -neutral multipole interaction is 9. M.R. Flannery, Phys. Rev. A 22 2408 (1980).
10. M. R. Flannery and D. Vrinceanu, Phys. Rev. Letts.
(Ze)Dn (Ze)Qn 85 (2000).
Vel (Ze; D, Q) = − 2
P1 (ŝ · R̂) + P2 (ŝ · R̂) 11. D. Vrinceanu and M.R. Flannery, Phys. Rev. A 6
R R3
  1053 (1999).
where Dn = rρ(r) dr and Qn = (3z 2 − r 2 )ρ(r) dr are 12. B. H. Bransden and M. R. C. McDowell, Charge
the permanent dipole and quadrupole moments of the Exchange and the Theory of Ion-Atom Collisions, (
neutral. The ion dipole - neutral dipole interaction is Clarendon Press, Oxford, 1992).
13. B. H. Bransden, Atomic Collision
Di Dn
Vel (Di ; Dn ) = − [2cos θi cos θn − Theory, (Benjamin-Cummings, Menlo Park, 1983),
R3 2nd ed.
sin θi sin θn cos(φi − φn )] 14. M. S. Child, Molecular Collision Theory, (Dover
Publications,Inc., New York, 1996).
where θi and θn are the angles made by the ionic and
15. P.G. Burke, in Atomic, Molecular and Opti-
molecular dipoles Di and Dn and the line R of centers
cal Physics Handbook , edited by Gordon W.
and φi and φn are the azimuthal angles of rotation about
Drake (American Institute of Physics Press, 1996),
the line of centers. The dipole - molecular quadrupole
Chap.45.
interaction is
16. Computational Atomic Physics, edited by Klaus
3Di Qn Bartschat (Springer, New York, 1996).
Vel (Di , Qn ) = [(3 cos2 θn − 1) cos θi − 17. I.E. McCarthy and E. Weigold, Electron-Atom Col-
2R4
2 sin θi sin θn cos θn cos(φi − φn )] lisions, (Cambridge University Press, Cambridge,
1995).
The Dispersion Interaction arises between the fluctu- 18. D.R. Bates in Quantum Theory I. Elements, edited
ating multipoles and the moments they induce and can by D. R. Bates (Academic Press, New York,
occur even between spherically symmetric ions and neu- 1961),Chap. 8.
trals. 19. D.R. Bates in Atomic and Molecular Processes,
 Encyclopedia of Chemical Physics and Physical Chemistry / 29

edited by D. R. Bates (Academic Press, New York, man (Wiley, New York, 1970).
1962),Chap. 14. Atomic and Molecular Processes, edited by D. R.
20. E.E. Nikitin, in Atomic, Molecular and Opti- Bates (Academic Press, New York, 1962).
cal Physics Handbook , edited by Gordon W. Quantum Theory I. Elements, edited by D. R. Bates
Drake (American Institute of Physics Press, 1996), (Academic Press, New York, 1961).
Chap.47. Electronic and Ionic Impact Phenomena, edited by
21. E. E. Nikitin and S. Ya. Umanskiı̆, Theory of Slow H. S. W. Massey, E. H. S. Burhop, and H. B. Gilbody
Atomic Collisions, (Springer-Verlag, Berlin, 1984). (Clarendon Press, Oxford, 1969-1974), Vols. 1–5.
22. K.J. McCann and M.R. Flannery, J.Chem.Phys. 12 Atomic Collisions: Electron and Photon Projectiles,
5275 (1978);63 4695 (1975). E. W. McDaniel, (Wiley, New York, 1989).
23. D.R. Bates and D.S.F. Crothers, Proc. Roy. Soc. A Atomic Collisions: Heavy Particle Projectiles, E. W.
315 465 (1970) McDaniel, J. B. A. Mitchell, and M. E. Rudd, (Wiley,
24. D.S.F. Crothers,Adv. Phys. 20 405 (1971) New York, 1993).
25. M. S. Child, Semiclassical Mechanics with Molecular
Applications, (Clarendon Press, Oxford, 1991).
26. N. Balakrishnan, R. C. Forrey and A. Dalgarno, 2.12 LIST OF KEYWORDS
Phys. Rev. Letts. 80 3224 (1998).
• Collisions
27. M.R. Flannery and K.J. McCann, in preparation.
28. M.R. Flannery and K.J. McCann, Phys. Rev. 9 1947 • Differential Cross Sections
(1974). • Integral Cross Sections
• Center of Mass
2.11 FURTHER READING • Elastic Scattering
• Inelastic Scattering
Atomic, Molecular and Optical Physics Handbook ,
• Reactive Scattering
edited by Gordon W. Drake (American Institute of
Physics Press, 1996). • Center-of-Mass to Laboratory Cross Section Conver-
Guide to Bibliographies, Books, Reviews and Com- sion
pendia of Data on Atomic Collisions, by E. W. McDaniel • Macroscopic Rate Coefficients
and E. J. Mansky in Advances in Atomic and Molecular
Physics, edited by B. Bederson and H. Walther Vol. 33, • Scattering Rate
389 (1994). • Energy Transfer Rate
Advances in Atomic and Molecular Physics, Vols. • Transport Cross Sections
1, 1965−1, 1973, edited by D.R. Bates and I. Estermann;
Vols. 10, 1974 − 25, 1988, edited by D.R. Bates and B. • Collision Integrals
Bederson. • Mobility Formula
Advances in Atomic Molecular and Optical Physics, • Quantal Transition Rates
Vols. 26, 1989 − 31, 1993, edited by D.R. Bates and B.
Bederson. Vol. 32, 1994, edited by B. Bederson and • Cross Sections
A. Dalgarno. Vols. 33, 1994 − 38, 1998, edited by B. • Interaction Matrix Element
Bederson and H. Walther. • Transition Operator
Atomic and Molecular Beam Methods, edited by G.
Scoles (Oxford University Press, New York, 1988). • Detailed Balance
Theory of Chemical Reaction Dynamics, Vols. 1-4, • Current
edited by M. Baer (CRC Press, Boca Raton, Florida, • Inelastic Cross Sections
1985).
Atom-Molecule Collision Theory: A Guide for the • Collision Strengths
Experimentalist, edited by R. B. Bernstein (Plenum • Excitation
Press, New York, 1979). • De-excitation
Dynamics of Molecular Collisions, Parts A and B ,
edited by W.H.Miller (Plenum Press, New York, 1976). • Dissociative Recombination
Case Studies in Atomic Collision Physics, edited by • Associative Ionization
E.W. McDaniel and M.R.C. Mc Dowell, (North-Holland, • Radiative Recombination
Amsterdam, Vol.1 (1969), Vol.2,(1972))
Ion-Molecule Reactions, edited by E. W. McDaniel, • Photoionization
V. Čermák, A. Dalgarno, E. E. Ferguson, and L. Fried- • Born Cross Sections
30 / Electron, Ion and Atom Scattering

• Fermi Golden Rules • Impact Parameter Method

• Electron - Atom Collisions • Landau - Zener Cross Section
• Ion- Atom Collisions • Averaged Potential
• Atom - Atom Collisions • Multichannel Eikonal Method
• Impulse Cross Sections • Long Range Interactions
• Atomic Form Factor
• Generalized Oscillator Strength
• Sum Rules
• Energy change moments
• Rotational Excitation
• Model Potentials
• Quantal Potential Scattering
• Partial Wave Expansion
• Scattering Amplitudes
• Integral cross sections
• Differential cross sections
• Legendre Polynomial Expansion
• Optical Theorem
• Levinson’s Theorem
• Partial Wave Expansions
• Born Phase Shifts
• Coulomb Scattering
• Collisions between Identical Particles
• Fermion Scattering
• Boson Scattering
• Coulomb scattering of two identical particles
• Scattering of Identical Atoms
• Quantal Inelastic Heavy-Particle Collisions
• Adiabatic Formulation
• Diabatic Formulation
• Central Field
• Two-State Treatment
• Distorted-Wave Approximation
• Born Approximation
• Exact Resonance
• Identical Nuclei
• Partial Wave Analysis
• Close Coupling Equations
• R-Matrix Method
• Electron-Molecule Collisions
• Semiclassical Inelastic Scattering
• Classical Path Theory