• Aucun résultat trouvé

Amplitude systems for spin-1/2 particles

N/A
N/A
Protected

Academic year: 2021

Partager "Amplitude systems for spin-1/2 particles"

Copied!
29
0
0

Texte intégral

(1)

HAL Id: jpa-00210988

https://hal.archives-ouvertes.fr/jpa-00210988

Submitted on 1 Jan 1989

HAL is a multi-disciplinary open access archive for the deposit and dissemination of sci- entific research documents, whether they are pub- lished or not. The documents may come from teaching and research institutions in France or abroad, or from public or private research centers.

L’archive ouverte pluridisciplinaire HAL, est destinée au dépôt et à la diffusion de documents scientifiques de niveau recherche, publiés ou non, émanant des établissements d’enseignement et de recherche français ou étrangers, des laboratoires publics ou privés.

Amplitude systems for spin-1/2 particles

Michael J. Moravcsik, Jutta Pauschenwein, Gary R. Goldstein

To cite this version:

Michael J. Moravcsik, Jutta Pauschenwein, Gary R. Goldstein. Amplitude systems for spin-1/2 par-

ticles. Journal de Physique, 1989, 50 (10), pp.1167-1194. �10.1051/jphys:0198900500100116700�. �jpa-

00210988�

(2)

Amplitude systems for spin-1/2 particles

Michael J. Moravcsik (1,*), Jutta Pauschenwein (1) and Gary R. Goldstein (2,**)

(1) Institut für theoretische Physik, Universität Graz, A-8010 Graz, Austria

(2) Department of Theoretical Physics, University of Oxford, Oxford 0X1 3NP, G.B.

(Reçu le 25 avril 1988, révisé le 6 décembre 1988, accepté le 13 janvier 1988)

Resume.

2014

On discute les propriétés de différents systèmes d’amplitudes qui sont utilisés pour la

description des réactions entre particules de spin 1/2. On compare les différentes représentations et

on présente les transformations entre les différents ensembles sous forme de matrices.

Abstract. 2014 The properties of various amplitude systems used to describe reactions involving spin-1/2 particles are described, the systems compared, and the transformation matrices among them tabulated.

Classification

Physics Abstracts

9nnn

1. Introduction.

The analysis of particle reactions (in nuclear- and high energy physics) is facilitated by describing these reactions in a way which explicitely displays the dependence of the dynamics

on all the quantum numbers and parameters which are firmly established. In particular, the

evidence is compelling that the structure of such reactions which pertains to spin and angular

momentum is rich and complex, and that its understanding is a necessary part of eventually discovering the fundamental dynamical laws of particles, and, specifically, that of strong interactions.

The characterization of this spin- and angular momentum structure can be done in at least two ways. The time-honored, traditional way has been in terms of a partial wave expansion of

the reaction amplitudes, which yields the phase shifts. This approach has brought us much insight for many reactions, both in nuclear and in particle physics. The approach, however,

also has some disadvantages which are gaining increasing importance nowadays. Some of

these can be summarized as follows :

a) at all but the lowest energies the partial wave description requires a very large number of parameters ;

b) in most, though not all, of the reactions the truncation of the partial wave expansion

cannot be done cleanly on physical grounds, and the mathematical criteria for doing the

truncation are not unambiguous either ;

,

Article published online by EDP Sciences and available at http://dx.doi.org/10.1051/jphys:0198900500100116700

(3)

c) while in some reactions the decomposition into definite angular momentum states appears to illuminate some significant features of the underlying dynamics, in most reactions

this does not appear to be the case ;

d) when there are large and numerous inelastic channels competing with the elastic channel

being described, unitarity (which is the basis for a phase shift description) becomes too complicated to be a simplifying feature in the partial wave description ;

e) in order to perform a reliable partial wave analysis, especially at higher energies, one

needs an extensive set of data in a wide range of the momentum transfer (or scattering angle).

The second way to describe particle reactions is in terms of spin amplitudes which are

functions of the kinematic parameters (energy and angle) of the reaction. While this approach (contrary to the partial wave expansion) does not connect data at different angles and thus

loses some unifying power, it shows many advantages in other respects. Some of these can be summarized as follows :

a) the number of parameters needed for the description of the reaction is independent of

the energy at which the analysis is carried out ;

b) data at single angles can thus be analyzed, something that is not possible in the partial

wave approach ;

c) there is a great variety of amplitude systems which allows one to choose the one most suited for the particular purpose at hand ;

d) dynamic models can be compared with phenomenological descriptions easier and in a

more definitive way ;

e) the analysis is independent of any inelastic channels that may be présent ;

f) specific kinematic configurations (e.g. forward- or backward reactions, or reactions at

90°) can be studied much easier in terms of amplitudes ;

g) the description requires no approximations like the angular momentum truncation

needed in the partial wave approach.

For these and other reasons, amplitude analysis has gained importance and popularity in

recent years.

A frequently occurring and important class of reactions is the one containing only spin-1/2 particles. Nucleon-nucleon and nucleon-antinucleon reactions are perhaps the most promi-

nent, as well as boson-nucleon and boson-hyperon reactions. In other instances, like in electron-proton, electron-electron, and electron-positron scattering, the reaction type is also the same but is further constrained by the known electromagnetic nature of the dynamics. In

any case, descriptions of reactions involving only spin-1/2 particles pervade the literature.

In this practice, a variety of different amplitude systems have been used. As remarked

earlier, they are not so much competitors but they rather complement each other, since for different purposes different amplitude systems turn out to be the best to use. This fact, together with the large number of specific analyses which have been carried out in the past in the various amplitude systems, suggests that it is likely that most of these systems will continue to be used in the future. As a result, it is useful to offer a comparative discussion of

them, and it is perhaps even more useful to display explicitely how one can make a transition

from one of these systems to another. That is the aim of the present paper.

2. Général properties of the amplitude systems.

The number of amplitudes needed for the description of a reaction is, of course, independent

of what amplitude system we use, and depends only on the values of the spins of the particles

in the process, and on what conservation laws constrain the reaction matrix. In this paper our

interest, for practical reasons, will be in reactions with both parity conservation and time

(4)

reversal invariance constraining the reaction matrix, that is, in the reactions of the type A + B --> A + B with parity conservation. The conceptual conclusions of our discussion can,

however, be carried over also to reactions which, for example, are not constrained by time

reversal invariance or are not even parity conserving.

With such constraints from parity conservation and time reversal invariance, for A with spin 1/2 and B with spin 0, the number of amplitudes is 2, while if both A and B have spins of 1/2, the number of amplitudes is 6. In what follows, we will specifically discuss only the

reaction with four spin-1/2 particles. The simpler reaction with only two spin-1/2 and two spin-

0 particles has obviously analogous properties on a reduced scale which can easily be deduced

from the present discussion.

The parity-conserving and time-reversal-constrained reactions with four spin-1/2 particles

fall in two additional subsets : one in which A

=

B (and the number of amplitudes reduces to 5), and one in which this is not the case. The classic example for the first is p-p scattering (or

p-p scattering if charge conjugation constraints are taken into account), while an example for

the second is n-p scattering (if charge independence is not taken into account). Because of the

difference in the number of amplitudes, and because of the somewhat different amplitude conventions, the two cases require somewhat separate treatment.

In both cases, however, one can divide the various amplitude systems used in the past into several classes and by several criteria.

One way to make the classification is according to the simplicity of the relationship between

the bilinear combination of amplitudes and the experimental observables. There are

amplitude systems which are « optimal » [1] in this respect in that for them the matrix that represents this relationship is as close to diagonal as possible. Other amplitude systems are non-optimal in this respect, in that for them the matrix is either not in the form of a string of

submatrices along the main diagonal, or it is but these submatrices are larger than what is

minimally required.

,

For the reactions under consideration in the present discussion, the difference between

optimal and non-optimal representations is small. To understand why, let us remember that in the optimal representations the M-matrix is spanned by basis matrices in each of which only

one element is non-zero, and the density matrices are spanned by basis matrices with only one

non-zero element along the diagonal (the other elements being zero) and by matrices with two

symmetrically placed off-diagonal elements being non-zero (and the other elements being zéro).

Now for spin-1/2 particles these matrices are so small (two-by-two) that the above optimal

basis matrices and the more usual non-optimal basis matrices (e.g. the Pauli matrices) differ only to a slight extent. As long as the three coordinate directions utilized in these basis matrices (e.g. the x, y, and z in the Pauli matrices) are chosen to be orthogonal to each other, the resulting non-optimal representation will differ from the optimal ones (as far as simplicity

is concerned) only slightly or not at all, especially when parity conservation and time reversal invariance are imposed also beside Lorentz invariance. Other advantages of the optimal

formalism (e.g. the flexibility in the choice of bases matrices, see the earlier discussion) will

however remain.

The other way of making a distinction among the various amplitude systems in use is according to whether they use space-fixed or particle-fixed coordinate systems. In the former,

the three coordinate directions are fixed once and for all with respect to the space in which the reaction takes place. They can, for example, be defined in terms of the two directions of, a) the incoming particle momentum, b) the normal to the reaction plane. Thereafter all

particles (that is, both the incoming and outgoing particles) use this fixed coordinate system

for the three directions appearing in the tensors describing them.

(5)

In the particle-fixed representations each particle has its own coordinate directions, adjusted, for example, to its own momentum direction. There may be constraints on the relative position of these directions between two different particles, resulting from constraints from conservation laws, but even so, all particles in the reaction are on an equal footing with respect to the definition of the coordinate directions. This is quite different from the situation described earlier for the space-fixed amplitude systems where the particle which is initially

used to define the coordinate system for all four particles is treated preferentially.

One obvious consequence of the space-fixed coordinate systems is that in it the relationship

of the final state particle momenta (if the system was fixed through an initial particle) to the

coordinate directions of this final state particle depends on the reaction angle, while this is not the case for the particle-fixed systems. Consequently the transformation matrices between

space-fixed amplitude systems and particle-fixed amplitude systems will contain the reaction

angle, while the transformation matrices between two space-fixed or two particle-fixed amplitude systems will not contain that angle. As we will see, one system (the singlet-triplet system) is neither pure space-fixed nor pure particle-fixed, and hence its relationship to all

other systems contains the reaction angle.

For many purposes particle-fixed amplitude systems are simpler and « more natural », and

hence many of the more modern systems are of such type. Among these are the whole class of

optimal formalisms as defined in reference [1], and hence also the helicity formalism which is

a special case of the optimal formalism. The so-called « transversity formalism » in the strict

sense of the word is also particle-fixed, namely if it is meant to be an optimal formalism with the quantization axes in the direction normal to the scattering plane. There are, however, also non-optimal formalisms, space-fixed, where the quantization axes are also in the direction normal to the reaction plane. An example would be a formalism which uses the Pauli spinors

as bases with the coordinate (usually called z) which is connected with the diagonal Pauli

matrix being identified with the direction normal to the reaction plane. One needs, therefore,

be careful with the terminology.

The various formalisms also differ from each other according to the properties they focus

on in their definitions. For example, the helicity formalism was defined because of its simple properties under coordinate transformations, and because a partial wave decomposition in it

appears simple if written in terms of the tabulated d-functions, whereas the corresponding decomposition, for example in some of the space-fixed amplitude systems, contains explicitely

many 9- j symbols and other purely geometrical factors which are hidden from the eye in the d- . functions occurring in the helicity formalism. On the other hand, the helicity formalism is not

particularly natural with regard to the parity constraints, and it does not appear to be

particularly natural either in describing the dynamical mechanisms occurring in strong interaction processes. Some of the space-fixed systems were introduced at a time when relativistic effects were small in the then available energy range in the reactions which were of interest at that time, and hence such systems, written in terms of quantities well known from non-relativistic quantum mechanics, were preferred.

Some other systems were specifically constructed to be simple in a particular dynamical

model. For example, the optimal « magic » amplitude system was defined specifically for the

purpose of ascertaining the extent to which one-particle-exchange (OPE) mçchanisms

dominate the reaction. Another example is the exchange amplitude system, originally

introduced during the Regge-era of particle physics, which appeared natural in that dynamical

framework. The planar-transverse system was, on the other hand, introduced not on a priori grounds, but only after an analysis of data, and its justification is in its exhibiting some striking

and simple features qf the data actually obtained by experiment.

Some types of systems show advantages even. in a purely phenomenological analysis of the

(6)

data, without regard to the dynamics. For example, the optimal transversity system is a natural one to use in such phenomenological analysis of any parity conserving reaction, since

its quantization direction (normal to the reaction plane) coincides with the preferential

direction in the symmetry operation of reflection which is the basis for parity conservation.

Thus we see indeed that there is a good reason to have such a multitude of different amplitude systems. In the next section, therefore, an overview of a list of these systems is given and their connections are specified.

3. Définitions and transformation matrices.

°

We will now define the various amplitude systems and then display their transformation matrices into each other. For some previous contributions to this task, see references [2, 3].

The systems we will consider are tabulated in table I. There are ten systems for the case with Lorentz invariance, parity conservation, time reversal invariance, and identical particles,

and seven systems for the case when the last (identical particle) constraint is not imposed.

NORMALIZATION. 2013 The normalization of the various sets of amplitudes is done in many different ways in the literature. In some cases the amplitudes themselves are dimensionless,

Table I.

z

The amplitude systems considered in this paper. The de finitions of the various amplitudes are given in the text.

Identical Particles

Non-identical particles

JOURNAL DE PHYSIQUE. - T. 50, N 10, 15 MAI 1989

(7)

and are normalized to some dimensionless number like 1 (but in some cases 1/2 or 2). In other

usage the amplitudes have a dimension of the square root of cross sections, that is, a

dimension of length. It would be impossible to cover all the possibilities which appear in the literature. Hence we concentrated in this paper on the relative normalizations of the various

amplitude systems. In table 1 we give, for each set, that sum of the bilinear combination of

amplitudes which should be taken to have the same value for all the different sets. This combination is the one that enters in the totally unpolarized differential cross section for that

amplitude system. This way the different amplitude systems can be compared, and it is this normalization which served as a basis for the transformation matrices given in this paper.

When then some amplitude system from a paper in the literature is used, the first task should be ascertain from that paper what the expression there is for the totally unpolarized

differential cross section, and then the amplitudes from there should be renormalized to agree with the normalization used in this paper. Then such amplitudes can be used in the transformation matrices given in this paper.

THE PROPERTIES OF THE AMPLITUDES SYSTEMS. - We will now briefly comment on each of

the nine amplitude systems and give their definitions.

1 ) The Saclay system [4]. - Named so because it appears in reference [2] with authors mainly

from Saclay. The M-matrix here is given by

where

This is for A different from B. For the identical particle case we have

This system uses the non-relativistic notation (which, however, as it has been shown [5], can

also handle the relativistic situation if interpreted the proper way). If the z direction (the one

for which the Pauli matrix is the diagonal one) is taken to be the normal to the reaction plane,

then this system is of the transversity type but not with a particle-fixed coordinate system (and

hence not a member of the set of optimal systems). The system is important because the

Saclay group, which defined it, has performed many experiments of the type under consideration which are then published within the framework of this formalism.

2) The Wolfenstein system [6].

-

Defined only for the identical particle case, in it the M-

matrix is given by

(8)

where S is the spin-singlet projection operator, T the spin-triplet projection operator, with the definitions

This is basically of the same type as the Saclay system, and is very simply related to it, with several of the amplitudes in the two systems being the same except for a numerical factor. Its

significance is historical, in that the first complete reaction analyses in the 1950’s were carried out using this amplitude system. It has also been used since.

3) The Hoshizaki system [7].

-

Also defined only for the identical particle case, in it the matrix is given by

This is even more similar to the Saclay system than the Wolfenstein system is. Indeed, the

difference is only in two of the amplitudes which, in the Hoshizaki system, are sums and differences of two amplitudes in the Saclay systems, the others being the same in the two systems except for factors of 2. The importance of this system is in the fact that Hoshizaki has

published a number of analyses of nucleon-nucleon scattering at energies higher than others,

and these analyses use this amplitude system.

The three above sets differ little from each other.

4) The Singlet-triplet system [2, 3].

-

This system is significantly different from the previous

ones (as well as from the upcoming ones) in that it focuses on a simple description of the two

f

particle systems occurring in the initial and final states. The M-matrix here has the following

form in the « singlet-triplet space » :

This is for A different from B ; for identical particle MST

=

0.

One then needs angular momentum addition geometry to relate these (larger number of)

matrix elements to the (smaller number of) amplitudes in the previous systems. As will be

seen, this ST system is not quite space-fixed (as the previous systems were), but not quite particle-fixed either (as the upcoming systems will be), and hence the transformation matrices between ST and any of the others will contain dependence on the reaction angle.

5) The Helicity system [8].

-

This is the first particle-fixed system that gained wide use. It

also featured easy Lorentz transformation properties which are perhaps nowadays not as important as they were at that time. It is the « natural » system for massless particles. There

are several conventions in use about some aspects of the definition of the helicity amplitudes,

as can be seen from references [8, 9]. In reference [2] the five amplitudes Ml, ...,

M5 are used. In this paper we use (a, b, c, d, e).

(9)

Table II.

-

The relative normalization of the various amplitude systems used in this paper.

Identical Particles

Non-identical particles

In reference [3] six helicity amplitudes are used, which are related to the ones used in this

paper by

(10)

The amplitudes used in this paper, in terms of the helicity values of particles C, A, D and B, respectively, in the reaction A + Bu C + D, are defined as follows :

The helicity system is a special case of the planar optimal systems, a family of systems which also includes the planar-transverse and the « magic » systems (see below).

6) The transversity system [10, 11].

-

This is a special case of the optimal systems which is particularly useful for the phenomenological analysis of parity-conserving reactions, since in it

the parity constraints on the amplitudes are simple, and since in it the simple polarization (with only one of the particles in the reaction being polarized) is given in terms of magnitude-

squares of reaction amplitudes, thus making the determination of these magnitudes easy.

In references [2] and [3] the transversity amplitudes Tl, ..., T6 are used. They are related to

the transversity amplitudes used in this paper by the correspondence

In particular, the former set is defined (see Ref. [15]), in terms of the spin projections of the particles C, A, D, B, in the reaction A + B --. C + D, as follows :

7) The Planar-transverse system [12]. - This system is a planar optimal system in which the

quantization axis is perpendicular to the helicity direction. There is growing evidence that, for

a variety of strong interaction processes, the reaction amplitudes have a strikingly simple

structure in this system, the relative phases among amplitudes being multiples of 90°. The amplitudes are defined analogously to the helicity amplitudes (see above).

8) The « Magic » system [13]. - This planar optimal system is used for one-particle-exchange

tests, since in it these tests have a particularly simple structure. It can be obtained from the

helicity system by a rotation of the quantization axis in the reaction plane by an angle which, however, depends on the reaction energy and reaction angle. For this reason the

transformation from the « magic » system to any other system is somewhat more involved, as

the values of the elements of the transformation matrices depend on reaction energy and reaction angle. Therefore, in this paper, we give only the transformation matrix between the

helicity system and the « magic system ». The transformation matrix between the « magic » system and any other system can be obtained straightforwardly by the multiplication of the

transformation matrix between « magic » and helicity with the transformation matrix between

helicity and that other system.

The « magic » frame amplitudes are defined in terms of the planar amplitudes for which

rotation angles in the plane are related to crossing angles (between s and t channels). For the

case of A + B the six magic amplitudes (a m, b i , c m, d m, e m, b2) are related to helicity

amplitudes via

(11)

where

When A and B are identical spin-1/2 particles, the six amplitudes reduce to five with

so that the transformation matrix becomes

The angles by which the quantization axes of the « magic » system are rotated with respect

to the helicity or particle momentum direction are given by

where s

=

(pA + PB)2 is the square of the center-of-mass energy, t

=

(PA’ - P A)2 is the square

of the four-momentum transfer, the primes refer to the outgoing particles, and mA and MB are the masses of particles A and B. When we have A

=

B, then the single angle of

rotation becomes

(12)

9) The exchange amplitude system.

-

This has its origin in the once popular Regge model,

since it explicitely emphasizes the quantum numbers of certain exchange processes.

10) The Arndt system [14].

-

This has been used by Arndt in his extensive analysis of

nucleon-nucleon elastic scattering. His program (SAID) provides spin amplitudes in this

form. The definitions of his amplitudes Hl, ..., HS are related to helicity amplitudes as

follows :

.

k... center-of-mass momentum.

It is useful to note that in reference [14] the fi amplitudes are related directly not to the helicity amplitudes but to the Saclay amplitudes.

11 ) The use of the transformation matrices.

-

In what now follows, we give the matrices which transform the amplitudes from one set to another. The notation is as follows : in order to obtain the amplitudes in system W from those in system V, one needs the two column

vectors (called Zv and Zw) of the five (or six) amplitudes in systems V and W, respectively,

and then the transformation matrix (five-by-five or six-by-six) Q y -+ W from the collection of such matrices to be given below. The amplitude vectors are given in table I, in row form for the sake of a more economical display in a printed table. The order of the amplitudes in these

vectors is, of course, the important piece of information. We thus have

The transformation matrices now follow.

(13)
(14)
(15)
(16)
(17)
(18)
(19)
(20)
(21)
(22)

JOURNAL DE PHYSIQUE. - T. 50, N’ 10, 15 MAI 1989

(23)
(24)
(25)
(26)
(27)
(28)

References

[1] GOLDSTEIN G. R. and MORAVCSIK M. J., Ann. Phys. N.Y. 98 (1976) 128.

[2] BYSTRICKY J., LEHAR F. and WINTERNITZ P., J. Phys. France 39 (1978) 1.

[3] LA FRANCE P. and WINTERNITZ P., J. Phys. France 41 (1980) 1391.

(29)

[4] OEHME R., Phys. Rev. 98 (1954) 147, 216.

PUZIKOV L. D., RYNDIN R. M. and SMORODINSKII Ya. A., Zh. Eksp. Teor. Fiz 32 (1957) 592.

LEHAR F. and WINTERNITZ P., Fortschr. Phys. 15 (1967) 495 ; reference [2], equation (2.1).

[5] STAPP H. P., Phys. Rev. 103 (1965) 425.

[6] WOLFENSTEIN L. and ASHKIN J., Phys. Rev. 85 (1952) 947 ; reference [2], equation (2.9).

[7] HOSHIZAKI N., Suppl. Prog. Theor. Phys. 42 (1968) 1, 107 ; reference [2], equation (2.6).

[8] JACOB M. and WICK G. C., Ann. Phys. N. Y. 7 (1959) 404 ; reference [2], equation (2.23).

[9] COHEN-TANNOUDJI G., MOREL A. and NAVELET H., Ann. Phys. N. Y. 46 (1968) 239.

[10] KOTANSKI A., Acta Phys. Pol. 29 (1966) 699 ; reference [2], equation (2.30).

[11] GOLDSTEIN G. R. and MORAVCSIK M. J., Ann. Phys. N. Y. 142 (1982) 219.

[12] GOLDSTEIN G. R. and MORAVCSIK M. J., Phys. Lett. B 102 (1981) 189.

[13] GOLDSTEIN G. R. and MORAVCSIK M. J., Phys. Rev. D 30 (1984) 55.

[14] ARNDT R., Phys. Rev. D 28 (1983) 97.

[15] GOLDSTEIN G. R. and MORAVCSIK M. J., J. Phys. Colloq. France 46 (1985) C2-251.

Références

Documents relatifs

We report mathematical results, which define a func- tion µ under which quantum selection takes place: as the spin grows, low-energy quantum states localize on the set of phase space

The temperature dependence of the longitudinal and trans- verse component of the magnetization moments m z 3 and m x 3 for the honeycomb lattice (z " 3) in zero transverse field

limit and reformulating perturbation theory in terms of resolvents we are able to obtain relativistic correction formulas (2.30), (2.32) for the eigenvalues of the

Annales de l’Institut Henri Poineare-Section A.. THE QUANTUM EQUIVALENCE PRINCIPLE, AND SPIN 1/2 MASSIVE PARTICLES 85. Substituting these equations in eq.. DIRAC’S EQUATION

predictive relativistic mechanics, that is, by fixing the reference frame (in our case by means of the center of mass condition). This statement

Let us now consider an isolated system of two EP-MD’s in interaction and analyze the « equations of motion )) given by Classical Field

- We develop the Predictive Relativistic Mechanics of isolated systems of particles with spin, using both the Manifestly Predictive.. Formalism and the Manifestly

Since the extremal symmetric states are tensorial infinite products of copies of a state (3.2.1), these states are invariant through the group of6. automorphisms