PHYSICAL REVIE% A VOLUME 14, NUMBER 6 DECEMBER 1976

High-momentum-transfer atomic collisions *

David Eimerl
Physical Dynamics, Incorporated,P. O, Box 556, La Jolla, California 92038
(Received 8 September 1976)
Recent results due to Kelsey for certain high-momentum-transfer collisions are generalized to all such
A
processes. simple model for scattering in this region is presented which provides considerable insight into
these processes. The model predicts that all processes in the regions where first Born terms fail are
proportional to the nuclear Rutherford-like scattering, and gives a simple formula for calculating the cross
section in terms of electronic form factors.

Recently there has been some interest'~ in col- mass, may change its energy by a large amount
lisions at high but nonrelativistic energies and and yet change its momentum by a small amount,
large momentum transfers. The interest has usually of order 1/a, . At large momentum trans-
focused on inelastic electron scattering off simple fers this is a negligible reaction on the nuclei.
atoms, where there is reason to believe that the Therefore one expects the motion of the nuclei to
'
first Born term fails in this region. Kelsey has be almost identical for both elastic and inelastic
shown' that for these processes the second Born processes; the dynamical role of the electrons is
term indeed dominates the first. It was found that irrelevant in determining the angle of scatter, As
the cross section at fixed energy does not drop far as an atomic electron is concerned it responds
rapidly with increasingly large momentum trans- to the changing fields during the collision but pro-
fer but remains proportional to a Rutherford-like vides a negligible reaction on the nuclei providing
scattering off the nucleus. The purpose of this the fields. Thus one may regard the nuclei as
comment is to point out that this feature is a gen- under going a Rutherford-like scattering, which
eral property of all atomic collisions at high mo- at high momentum transfers is independent of the
mentum transfer, and to develop a simple model screening lengths, and at the same time providing
which circumvents the difficulties of computing fields that perturb the atomic electrons. Thus
higher-order effects. Such a model facilitates the these fields may be obtained using the motion of
evaluation and comparison of data and aids in the nuclei to provide a specified source for the
identifying atomic phenomena specific to this re- perturbing fields. This picture leads toa factor-
gion. To the author's knowledge, no such model ized scattering amplitude
has been discussed in the literature.
The present model is based on the following
semiclassical ideas. It is well known that the where F~ describes the nuclear amplitude and A,
elastic collisions of atoms or ions at high energies is computed using perturbation theory to describe
are accurately described by the screened Coulomb the electronic transition produced by the nuclear
potential of the nuclei, away from the forward di- fields. The basic approximation here is ne- the
rection. The distance of closest approach in an
proachh.
glect of the electronic process on
reaction of the
elastic collision may be much smaller than the the nuclei, thus allowing the semiclassical ap-
atomic dimension, which shows that the electrons proach of treating the electronic transition produc-
play no dynamical role, merely passively provid- ing fields using a specified c-number source. The
ing a screen for the internuclear interaction. The validity of the model depends on a very close nu-
scattering is due entirely to the forces felt by the clear collision, and therefore requires high
nuclei as they reach the distance of closest ap- energy and large momentum transfer. The basic
idea is similar to the straight-line approximation
In order to generalize this to inelastic processes for low-energy distant collisions between a
at high momentum transfer, it is argued that dur- charged particle and an atom or ion. The physical
ing such a collision an electron in an atom typical- basis is, however, completely different.
ly changes its momentum by an amount of the or- The cross section may be obtained dixectly from
der of the normal momentum of a bound electron. Eq. (1):
This holds also for ionization reactions where the
probability of producing a high-energy electron is
4nM~Z, Z.
2 , Ni
(,)
small compared to a lower-energy electron. The
point is that an electron, because of its negligible where ~lI~ is the reduced mass and f„ is the nu-

2061
 DAVID EIMERL

clear strong-interaction amplitude, which becomes may be neglected, and the amplitude for the elec-
important at large q'. In first Born approximation trons to be in a final state s is
P is simply a product of atomic form factors, and
depends on q'. At high momentum transfer the
))
Horn approximation to P drops rapidly in q' for C2 g +~ e df ci(E& E&-)t (~(2 Igil)j(2 I~{2
inelastic processes, remaining finite and equal
to unity only for truly elastic processes. We where the space components of the electromag-
shall show that the present model for P gives netic potential have been dropped. The electronic
unity for elastic processes and a quantity of order transition probability P is then given by
(uc/u)' for inelastic processes. The model gives C(2)c(1)~~
P in terms of electronic form factors; we now I

turn to its evaluation.

The current of an atom scattering from a mo- It is emphasized that only the electronic contri-
mentum p to p', is well known from the theory of bution to jo~'~ is to be included in Eq. (10). The nu-
bremsstrahlung": clear term is to be excluded since it represents a
higher-order approximation to the nuclear scatter-
ing amplitude, which in this analysis is a correc-
tion to da„/did and not a contribution to P IThi. s
is already implicit in Eq. (8) where only electronic
coordinates appear. J To evaluate the coefficients
C„write
~V'0'*') J+O'I&, ="s'a/I2
Ie'" I'
The index p. runs from 0 to 3, and all vectors in
these expressions are four-vectors. The current to obtain from (10)
is proportional to a form factor, which is the
(screened) nuclear charge, and has the limits

k=E-E.
E, (0) Z, N,
= - The dependence on k, is very weak, and neglecting
where N is the number of electrons. In this semi- it gives a simple perspicuous result for C, :
classical approach the nuclear charge is the dom- v„+ P(v) +, (v)-
inant contributor to electronic transitions. The ~c 1 1 "dip, '„dQ,
current produces an electromagnetic field of a
Lienard-Wiechert type, ci'& = ——+ —, — ",
—
A„"(0)=j „'(0)/(0' f e ),
which produces transitions in the other atom. Together with Eqs. (11) and (2) this completes the
These may be computed by simple perturbation model.
theory. The wave function for atom 2 may be In the case of the elastic scattering of two ions
written as follows
or an ion and a charged particle, the integral in

f, "'(~„f) = g c~"(f, ~)q~'~(x,

& exp
).
i (g„+—,~&~, tP)t
-
— i w.,v x,
(15) diverges. However, this occurs only for the elastic
case, in which the integral is formally of order ~
compared to the constant term. This is
the same phenomenon
approximation and is actually
as occurs in second Born
related to the charge
re nor malization of quantum electrodynamics, '
which possesses an infrared divergence of this
where v is the velocity of the atom. If very large type. Consequently, it is permissible to absorb
momentum transfers are excluded, namely that divergence into the charge renormalization
constants; the end result is simply to strike out
&&
gu, 3 M2/ill„, the integral in Eq. (15) and replace it by a finite
then the velocity dependence of the coefficients C„ quantity of order a, which can now be neglected.
14 HIGH-MOM ENTUM- T RAN SF E R ATOM IC COL LI SION S 2063

Then we obtain
27T O 0 V1 '~)1 41T
CI') =1
st i F'
and for
— — (k) where the refers to the final state of the atom
nc 1 1 "d& or ion.
C~" = — —+ —, F'(k)
F~' The application of the results to electron scatter-
ing is weaker in that the arguments based on the
(17) inertia of the atomic electrons are weaker. The
The model is clearly very simple to use and al-
assumption of neglecting their reaction on the
impinging electron is nonetheless expected to be
lows the immediate comparison of several excita-
tion processes, since the final electronic state quite good. The reason is that atomic electrons
appears only in F„.
Rather than attempting to typically change the projectile momentum by some-
thing of the order of 1/a„, which is small com-
tabulate the cross sections for a large number of
processes, it is preferable to have the model in pared to the momentum transfer between the pro-
its present analytic form so that any process of jectile and the nucleus. The results of applying
this model to electron scattering off simple atoms
interest may be immediately calculated by the
give results very close to those of Kelsey, ' and
reader as required. The F„
form factors may
pr ovide ins ight into his results.
be obtained from data at lower energies or from
In conclusion, a general, simple model of large
simple models. The F form factors describing
momentum transfer processes has been presented.
the screened nuclear charge may be taken to be
The model predicts that all such processes are
the ground-state form factor, or a screened po-
proportional to the nuclear Rutherford-like cross
tential where the screening length is some
section, where the coefficient is independent of
weighted average of the initial and final-state
screening lengths. This is expected to be suffi- angle l(cf. (18)j and of order (o.c/v)'. Any devia-
tion from such behavior would indicate new and
ciently precise if these screening lengths are not
interesting phenomena in this region.
very different. If ionization occurs, then this ap-
proximation may not be sufficient. In that case, The author is grateful for helpful conversations
Eq. (17) may be replaced by the following: with Professor M. Rotenberg and Dr. G. Gillespie.

*This research was sponsored by the Department of the E. J. Kelsey and J. Macek, Phys. Rev. A 12 1127 (1975).
Army, Ballistic Missile Defense, Advanced Technology C. Quigg and C. J. Joachim, Rev. Mod. Phys. 46 279
Center, Huntsville, Ala. , under Contract number (1974).
DASG 60-75-0080. J. D. Jackson, Classical Electrodynamics (Wiley, New
'S. Geltman and M. B. Hidalgo, J. Phys.B 4 1299 (1971). York, 1962), Chap. 14.
E. J. Kelsey, Phys. Rev. A 14, 56 (1976). 'J. D. Bjorken and S. D. Drell, Relativistic Quantum
B. R. Junker, Phys. Rev. A 11, 1552 (1974). Iields (McGraw-Hill, New York, 1965), Chap. 19.