Quantization of spherically symmetric solution of SU(3)

Yang-Mills theory

∗
V. Dzhunushaliev
Theor. Physics Dept., Kyrgyz State National University, 720024, Bishkek, Kyrgyzstan

†
arXiv:hep-th/9810094v1 14 Oct 1998

D. Singleton
Dept. of Physics, CSU Fresno, 2345 East San Ramon Ave., M/S 37, Fresno, CA 93740-8031

(February 5, 2018)

Abstract

A recent investigation of the SU(3) Yang-Mills field equations found sev-

eral classical solutions which exhibited a type of confinement due to gauge

fields which increased without bound as r → ∞. This increase of the gauge

fields gave these solutions an infinite field energy, raising questions about

their physical significance. In this paper we apply some ideas of Heisenberg

about the quantization of strongly interacting, non-linear fields to this clas-

sical solution and find that at large r this quantization procedure softens

the unphysical behaviour of the classical solution, while the interesting short

distance behaviour is still maintained. This quantization procedure may pro-

vide a general method for approximating the quantum corrections to certain

classical field configurations.

Pacs [Link]

Typeset using REVTEX

∗ E-mail address: dzhun@[Link]

† E-mail address : dasingle@[Link]

1
 I. INTRODUCTION

Recently [1] several classical solutions to SU(3) Yang-Mills theory were discussed, which

possessed either spherical or cylindrical symmetry. These solutions had gauge fields which
tended toward ∞ at large distances, leading to a type of confining behaviour if one considered
these solutions as background fields in which some test particle moved. These increasing
gauge fields also led these solutions to the undesired property of having infinite field en-

ergy. One way in which these classical field configurations might nevertheless have some
physical importance is if the quantization of these solutions reduced or eliminated the bad
long distance behaviour. While perturbative quantization techniques work well for weakly
interacting field theories such as QED (or QCD in the high energy limit), they are not
useful when dealing with strongly interacting field theories. In Ref. [2] we applied some

ideas of Heisenberg’s concerning the quantization of strongly interacting, non-linear fields

[3] to the cylindrical solution discussed in Ref. [1], and found that under certain assumptions
the bad long distance behaviour of this solution was eliminated. Here we apply the same
procedure to the spherical symmetric solution and show that again the bad long distance

behaviour is eliminated. In addition to the specific benefit that the Heisenberg quantization
method gives to the infinite energy solutions discussed here and in Ref. [2], it may provide
some general procedure for approximating the quantum corrections to certain classical field
configurations.

II. SPHERICALLY SYMMETRIC ANSATZ

We will briefly review the derivation and discuss some aspects of the spherically sym-
metric solution. The ansatz for the SU(3) gauge field we take as in [4] [5] [6]:

2ϕ(r)  2 1 a a  i j
a x x

5 7
A0 = λ x − λ y + λ z + λ λ ij + λ ji w(r), (1a)
ir 2 2 r2
 xj  l
j x
 
Aai = λaij − λaji (f (r) − 1) + λ a
jk ǫilj x k
+ ǫ ilk x v(r), (1b)
ir 2 r3

2
here λa are the Gell - Mann matrices; a = 1, 2, . . . , 8 is a color index; the Latin indices
i, j, k, l = 1, 2, 3 are the space indices; i2 = −1; r, θ, φ are the usual spherically coordinates.
Substituting the ansatz of Eqs. (1) (with f = φ = 0) into the Yang - Mills equations

1 √ 
√ ∂µ −gF aµ ν + f abc F bµ ν Acµ = 0, (2)
−g

yields the following complex set of coupled, non-linear differential equations [4]

 
r 2 f ′′ = f 3 − f + 7f v 2 + 2vwϕ − f w 2 + ϕ2 , (3a)
 
r 2 v ′′ = v 3 − v + 7vf 2 + 2f wϕ − v w 2 + ϕ2 (3b)
 
r 2 w ′′ = 6w f 2 + v 2 − 12f vϕ, (3c)
 
r 2 ϕ′′ = 2ϕ f 2 + v 2 − 4f vw. (3d)

For the solution with increasing gauge fields we specialized by taking f = ϕ = 0 (the case

where v = w = 0 is similiar) gives the following set of non-linear coupled equations

r 2 v ′′ = v 3 − v − vw 2 , (4a)

r 2 w ′′ = 6wv 2. (4b)

Near r = 0 we took the series expansion form for v and w as

r2
v = 1 + v2 + ..., (5a)
2!
r3
w = w3 + ... (5b)
3!

where v2 , w3 were constants which determined the initial conditions on v and w as in the
last section. In the asymptotic limit r → ∞ the form of the solutions to Eqs. (4) approaches

the form

v ≈ A sin (xα + φ0 ) , (6a)

α − 1 cos (2xα + 2φ0 )
" #
w ≈ ± αxα + , (6b)
4 xα
3A2 = α(α − 1). (6c)

3
where x = r/r0 is a dimensionless radius and r0 , φ0 , and A are constants. The second,
strongly oscillating term in w(r) is kept since it contributes to the asymptotic behaviour
of w ′′ . We did not find an analytical solution for the system of Eqs. (4), but it is straight

forward to solve these equations numerically with any standard differential equation package
such as that available in Mathematica [7]. Fig. 1 shows a representative solution to Eqs. (4).
The exponent of the power law increase of w (which is represented by α in the asymptotic
expressions) depended on the initial conditions, which were determined by the constants

v2 , w3. Generally the exponent α would decrease from a value in the range 2 − 3 to a value
in the range 1.2 − 1.8 for a wide range of initial conditions. This behaviour can be seen
in the Log(w) − Log(x) plot in Fig. 2. Although, these classical gauge fields weakened
slightly as r increased, they still diverged as r → ∞. Due to this feature of the ansatz

function w the time part of the gauge field grew without bound as r → ∞, leading to both
a classical type of confinement (a test particle placed in the background field of this solution
would not be able to escape to ∞) and an undesired infinite field energy for this solution.
Various phenomenological studies of quarkonia bound states use such increasing potentials

to study the spectrum of the bound state [8] although usually the potential is taken to
increase linearly. It should be mentioned that the asymptotic form of the classical solution
given in Eqs. (6) are expected to be altered by the quantum corrections. The classical,
short distance behaviour, as given in Fig. 1, should be roughly correct, since the pure gauge
SU(3) theory that we are considering is asymptotically free.

This “bunker” solution has “magnetic” and “electric” fields associated with it. Using
the ansatz for Aµ from Eq. (1) these “magnetic” and “electric” fields have the following
proportionalities

v2 − 1
Hra ∝ , Hϕa ∝ v ′ , Hθa ∝ v ′ , (7a)
r2
rw ′ − w vw vw
Era ∝ , Eϕa ∝ , Eθa ∝ , (7b)
r2 r r

here for Era , Hθa , and Hϕa the color index a = 1, 3, 4, 6, 8 and for Hra , Eθa and Eϕa a = 2, 5, 7.
The asymptotic behaviour of Hϕa , Hθa and Eϕa , Eθa is dominated by the strongly oscillating

4
function v(r). It may be postulated that quantum corrections to this strongly ocsillating
solution would tend to smooth it out so that it would not play a significant role in the large
r limit. From Eqs. (7) and the asymptotic form of v(r), w(r) the radial components of the

“magnetic” and “electric” have the following asymptotic behaviour

1 1
Hra ∝ , Era ∝ . (8)
r2 r 2−α

where the strongly oscillating portion of Hra is assumed not to contribute in the limit of

large r due to smoothing by quantum corrections. The radial “electric” field falls off slower
than 1/r 2 (since α > 1) indicating the presence of a confining potential. The 1/r 2 fall off
of Hra indicates that this solution carries a “magnetic” charge. This was also true for the
simple solutions discussed in Refs. [4] [5]. This leads to the result that if a test is placed

in the background field of the bunker solution, the composite system will have unusal spin
properties (i.e. if the test particle is a boson the system will behave as a fermion, and if the
test particle is a fermion the system will behave as a boson). This is the spin from isospin
mechanism [9].

By examining the classical SU(3) field equations of Eqs. (4) we have found field config-
urations which led to a classical confining behaviour, and which has some similarities with
certain phenomenological models used to study heavy quark bound states. The most signif-
icant draw back of the present solutions is that it has infinite field energy. The asymptotic
form of the energy density goes as
!2
v ′2 2 w′ w v2w2 2  2 2 2 α2 (α − 1)(3α − 1)
E ∝4 2 + − 2 +4 4 + 4 v −1 ≈ (9)
r 3 r r r r 3 x4−2α

Since we found α > 1 this energy density will yield an infinite field energy when integrated
over all space. This can be compared with the finite field energy monopole [10] and dyon

solutions [11] [12].

5
 III. QUANTIZATION OF THE “BUNKER” SOLUTION

Although the classical confining behaviour of this “bunker” solution may seem interest-

ing due to its similiarity with certain phenomenological potentials, the infinite field energy
discussed at the end of the previous section strongly argues against the physical importance
of this solution. One possible escape from this conclusion is if quantum effects weakened
or removed the bad long distance behaviour of these solutions. However, strongly inter-

acting, non-linear theories are notoriously hard to quantize. In order to take into account
the quantum effects on the bunker solution we employ a method used by Heisenberg [3] in
attempts to quantize the non-linear Dirac equation. By applying the dynamical equation of
motion (in Heisenberg’s case the non-linear Dirac equation) to an n-point Green’s function,
Gn , one would arrive at an equation relating Gn to higher order Green’s functions (Gn+1 for

example). Then applying the dynamical equations of motion to the higher Green’s functions
one would get equations relating these higher Green’s functions to even larger order Green’s
function. Continuing in this way one arrived at an infinite set of differential equations relat-
ing Green’s functions of all orders. To handle this Heisenberg employed the Tamm-Dankoff

method whereby he only considered Green’s functions up to some order thus cutting off the
infinite set of equations. Here we will employ a similiar method to the “bunker” solution
in terms of the ansatz functions v and w. Previously [2] we used this method on an infi-
nite energy, string-like classical solution to the SU(3) equations. For more details on the

application of the Heisenberg method to such classical solutions we refer the reader to this
article.
In order to use Heisenberg’s quantization method on the nonlinear equations we make
the following assumptions :
1. The degrees of freedom relevant for studying the “bunker” solution (both classically

and also quantum mechanically) are given entirely by the two ansatz functions w, v. No
other degrees of freedom arise through the quantization process.
2. From Eq. (6b) and Fig. 1 w is a smoothly varying function for large x, while v is

6
strongly ocsillating. Thus we take w(x) to be almost a classical degree of freedom while v(x)
is treated as a fully quantum mechanical degree of freedom. One might think that in this
way only the behaviour of v would change while w stayed the same. However since w and

v are coupled via the equations of motion we find that both functions are modified.
To begin we replace the ansatz functions by operators ŵ(x), v̂(x).

x2 v̂ ′′ = v̂ 3 − v̂ − v̂ ŵ 2 (10a)

x2 ŵ ′′ = 6ŵv̂ 2 (10b)

here the prime denotes a derivative with respect to x. Taking into account assumption (2)
we let ŵ → w become just a classical function again, and replace v̂ 2 in Eq. (10b) by its

expectation value to arrive at

x2 v̂ ′′ = v̂ 3 − v̂ − v̂w 2 (11a)

x2 w ′′ = 6whv 2i (11b)

where the expectation value hv̂ 2 i is taken with respect to fluctuations of v around the classical
solution of Eq. (6) or Fig. 1 (i.e. hv̂ 2 i = DveiS/h̄ v 2 where S is the action and Dv is a path
R

integral measure over all possible configurations of v). If we took the expectation value of

Eq. (11a) we would almost have a closed system of differential equations relating w and
hv̂i. The hv̂ 2 i term from Eq. (11b) and the hv̂ 3 i term from Eq. (11a) prevent the equations
from being closed. Applying the operation x2 ∂ 2 /∂x2 to the operator v̂ 2 and using Eq. (11a)
yields

x2 (v̂ 2 )′′ = 2v̂ 2 (v̂ 2 − 1 − w) + 2x2 (v̂ ′ )2 (12)

If we took the expectation of the above equation with respect to fluctuations in the ansatz

function operator v̂ 2 , and combined this with Eq. (11b) we would almost have a closed
system for determining w and v̂ 2 except for the h(v̂ ′)2 i term which comes from the last term
on the right hand side of Eq. (12). Continuing in this way one could obtain an infinite
set of equations for various powers of the ansatz function operator (i.e. v̂ n ). These higher

7
order equations never close. To deal with this problem we follow Heisenberg, and make some
assumption that effectively cuts off the system of equations at some finite order. By taking
the expectation of Eq. (12) and further by assuming that

hv̂ 2 i − v02
h(v̂ ′ )2 i = (13)
x2

we arrive at the closed system of equations from Eqs. (11b) (12)

x2 hv̂ 2 i′′ = 2hv̂ 2 i2 − 2hv̂ 2 iw − 2v02 (14a)

x2 w ′′ = 6whv̂ 2i (14b)

By making the assumption of Eq. (13) we have simplified Eqs. (11b) (12) to the closed

system given by Eqs. (14a) (14b). It is straightforward to show that in the limit x → ∞

a
hv̂ 2 i = v02 + (15a)
xα
b
w= (15b)
xα

solves Eqs. (14a) (14b) provided that v02 = 1, b = −a and α = 2, −3. In order for hv̂ 2 i
and w to have acceptable behaviour at x → ∞ we take the α = 2 solution. Substituting
the above expressions for hv̂ 2i with v0 = +1 and w back into Eq. (11a), and assuming that
hv̂ 3 i = hv̂ihv̂ 2 i gives the following equations for hv̂i in the x → ∞ limit

a
x2 hv̂i′′ = hv̂i(hv̂ 2i − 1) = hv̂i (16)
x2

Eq. (16) is solved by in the x → ∞ limit

a
 
hv̂i = ± 1 + 2 (17)
6x

Eq. (17) together with Eqs. (15a) (15b) provide information on the behaviour of the

“classical” ansatz function, w, and the “quantum” ansatz function, v, via hv̂i and hv̂ 2 i. The
main point of interest is that after applying the Heisenberg-like quantization procedure to
the classical solution of Fig. 1, the infinite increase of the ansatz function, w, has changed
to an acceptable asymptotic behaviour (i.e. one that leads to a finite field energy). By

8
replacing the v 2 and (v ′ )2 terms in Eq. (9) with hv̂ 2 i and (hv̂i′ )2 - from Eqs. (15a) and (17)
- respectively, and also using w from eq. (15b) we find that the field energy density of the
quantized “bunker” solution takes the form

a2
E∝ (18)
x8

in the limit in which quantum fluctuations become important (i.e. for non-Abelian theories
which exhibit asymptotic freedom this means in the low energy or x → ∞ range) the energy

density goes from the form given in Eq. (9) to that given in Eq. (18). This can be seen
to give a finite field energy. In the high energy or short distance regime we assume that
the fields approach the classical configuration of Figure 1 due to asymptotic freedom. This
classical configuration is well behaved at x = 0, but would yield an infinite field energy due

to its divergence as x → ∞. In the long distance or low energy limit the energy density
should go over into the form given in Eq. (18) which would then result in a finite field energy
for this configuration, since the integral of Eq. (18) over the large x region, where it is valid,
would give a finite field energy.

Finally by using w, hv̂i, and hv̂ 2 i in the expressions for E and H given in Eq. (7) we
find that the radial fields (Er , Hr ) go like a/r 4 while the angular fields (Eθ,φ , Hθ,φ ) go like
a/r 3 . Thus the quantization procedure outlined above modifies the undesirable long distance
behaviour of the “electric” and “magnetic” fields as well as the energy density.

IV. DISCUSSION

In this paper we reviewed a certain classical field configuration for an SU(3) gauge theory.
Near the origin the field configurations were finite, but as r → ∞ the fields diverged (see
Figs. 1 ,2). This increasing field strength led to a classical type of confinement in that
a test particle placed in the background field of this solution would not be able to escape

to ∞. Unfortunately, this diverging of the field as r → ∞ also led to this configuration

having an infinite field energy. Previously it was suggested that quantum effects might

9
soften or eliminate this bad long distance behaviour. By applying a method similiar to that
Heisenberg used in quantizing the non-linear Dirac equation we find that the long distance
behaviour is changed so as to give finite field energy. At short distances the fields should

approach the classical configuration of Fig. 1 from the asymptotic freedom of the SU(3)
gauge theory. This classical solution has the good features of not being divergent at r = 0
and in some limited region around r = 0 the fields increase in a way similiar to that found in
some phenomenological models of confinement. At long distances the fields should approach

the configuration given by Eqs. (15a) (15b) (17) where the quantum effects have eliminated
the divergence of the fields and field energy density as r → ∞.

10
 REFERENCES

[1] V. Dzhunushaliev, “Confining solutions of SU(3) Yang-Mills theory”, hep-th 9611096

[2] V. Dzhunushaliev and D. Singleton, “Quantization of strongly interacting fields”, hep-

th/9806073

[3] W. Heisenberg, Nachr. Akad. Wiss. Göttingen, N8, 111 (1953); W. Heisenberg, Zs.
Naturforsch., 9a, 292 (1954); W. Heisenberg, F. Kortel and H. Mütter, Zs. Natur-
forsch., 10a, 425 (1955); W. Heisenberg, Zs. für Phys., 144, 1 (1956); P. Askali and
W. Heisenberg, Zs. Naturforsc., 12a, 177 (1957); W. Heisenberg, Nucl. Phys., 4, 532

(1957); W. Heisenberg, Rev. Mod. Phys., 29, 269 (1957)

[4] W.J. Marciano, H. Pagels, Phys. Rev. D12, 1093 (1975)

[5] Z. Horvath and L. Palla, Phys. Rev. D14, 1711 (1976)

[6] D.V. Gal’tsov and [Link], [Link], B274, 173(1992).

[7] S. Wolfram, Mathematica 2nd Ed., (Addison-Wesley Publishing Company, 1991), p. 829

[8] E. Eichten et. al., Phys. Rev. D17, 3090 (1978)

[9] R. Jackiw and C. Rebbi, Phys. Rev. Lett. 36, 1116 (1976); P. Hasenfrantz and G. ’t
Hooft, Phys. Rev. Lett. 36, 1119 (1976)

[10] G. ’t Hooft, Nucl. Phys. B 79, 276 (1974); A.M. Polyakov, Sov. Phys.-JETP 41, 988

(1975)

[11] M.K. Prasad and C.M. Sommerfield, Phys. Rev. Lett. 35, 760 (1975); E.B. Bogomolnyi,

Sov. J. Nucl. Phys. 24, 449 (1976)

[12] B. Julia and A. Zee, Phys. Rev. D11, 2227 (1975)

11
 List of figure captions
Fig.1. The w(x) confining function, and the v(x) oscillating function of the SU(3) bunker
solution. The initial conditions for this particular solution were v2 = 0.1, w3 = 2.0, and

xi = 0.001.

Fig.2. A plot of Log(w) − Log(x) of the solution from Fig. 2 showing the different power
law behaviour in the small x and large x regions.

12