It is remarkable that quantum chromodynamics Gra95that describes all phenomena related to strongly interacting
particles, can be constructed in full analogy to the QED.
The ``only'' difference is a more complicated gauge group,
SU(3) instead of U(1). The complete QCD Lagrangian density reads
Dirac four-spinors correspond to quark fields. Compared to the electron four-spinors discussed in Sec. 2.2, they are richer in two aspects. First, each of them appears in three variants, red, blue, and green. These colors are numbered by the quark-color index corresponding to the dimensions 33 of the matrices. Traditionally they are not explicitly shown in the Lagrangian density (22), so we should, in fact, think about as 12-component spinors. One should not be confused by the fact that there are three colors of quarks, and eight colors of gluons - in fact, here the "visual" representation simply breaks down, and the colors of gluons have nothing to do with red, blue, and green of quarks. In reality, quarks and gluons are numbered by the indices of the corresponding SU(3) representations: three-dimensional spinor representation for quarks, and eight-dimensional vector representation for gluons.
Second, there is not one, but six different quark fields, for =1,...,6. These are called quark flavors, and are usually denoted by names: up, down, charm, strange, top, and bottom, see Fig. 1. For nuclear structure physics, essential rôle is played only by the up and down quarks that are constituents of neutrons and protons. So in most applications of the QCD to nuclear structure, we can limit the QCD Lagrangian density to two flavors only, =1, 2.
The first term in the QCD Lagrangian density (21) describes the free gluon fields defined by eight four-potentials . One can say that instead of one photon of the QED, that transmits the electromagnetic interaction, we have eight gluons that transmit the strong interaction, see Fig. 2.
The gluon field tensors
are defined as
The last term in (21) describes the six free quarks of masses
at rest. This does not mean that isolated quarks can exist
in Nature, be accelerated, and have their masses measured by
their inertia with respect to acceleration. Each free quark obeys the
same Dirac equation as the electron in QED. The Dirac equation is
given by the last term and the -term in Eq. (21). Quarks couple to gluons through the color currents,
Consequences of the gluon charges are dramatic. Namely, the force carriers now exert the same force as the force they transmit. Moreover, sources of the electromagnetic field depend on currents (20) that involve a small parameter - the electron charge, while gluons constitute sources of the color field without any small parameter. Gluons are not only color-charged, but they also produce very strong color fields.
Let us now consider empty space. In a quantum field theory, we cannot just say that the ground state of the empty space is the state with no quanta - we have to solve the proper field equations, with proper boundary conditions, and determine what is the state of the field. Such a state may or may not contain quanta. In particular, whenever the space has a boundary, the ground state of the field does contain quanta - this fact is called the vacuum polarization effect.
In QED, this is a very well known, and experimentally verified effect. For example, two conducting parallel plates attract each other, even if they are not charged and placed in otherwise empty space (this is called the Casimir effect Plu86). One can understand this attraction very easily. Namely, the vacuum fluctuations of the electron field may create in an empty space virtual electron-positron pairs. These charged particles induce virtual polarization charges in the conducting plates (it means virtual photons are created, travel to plates, and reflect from them). Hence, the plates become virtually charged, and attract one another during a short time when the existence of the virtual charges, and virtual photons, is allowed by the Heisenberg principle. All in all, a net attractive force between plates appears.
In QED such effects are extremely weak, because the electron has a small charge and a non-zero rest mass. On the other hand, the QCD gluons are massless, and their strong interaction is not damped by a small parameter. As a result, the QCD vacuum polarization effect is extremely strong, and the empty space is not empty at all - it must contain a soup of spontaneously appearing, interacting, and disappearing gluons. Moreover, in the soup there also must be pairs of virtual quark-antiquark pairs that are also color-charged, and emit and absorb more virtual gluons. It turns out that the QCD ground state of an "empty" space is an extremely complicated object. At present, we do not have any glimpse of a possibility to find the vacuum wave function analytically. Some ideas of what happens are provided by the QCD lattice calculations, see e.g. Ref. Lei00, in which the gluon and quark fields are discretized on a four-dimensional lattice of space-time points, and the differential field equations are transformed into finite-difference equations solvable on a computer.
It is now obvious that one cannot expect other solutions of the QCD fields to be any simpler. In particular, solutions corresponding to isolated quarks simply do not exist. One can say that an isolated quark would create so many gluons around it that the complete wave function had not been normalizable. Solutions for quark-antiquark pairs, and for triples of quarks, do exist (we do exist after all - the triples of quarks are nucleons our bodies are built of), but are even more complicated to obtain, even within the QCD lattice calculations. There is no hope, neither there is any reason, to describe composite objects like mesons or nucleons directly from quarks and gluons. This is especially true when we want to use these composite objects to build the next-generation composite objects like nuclei.
Here we arrive at the leading idea of our physical description of the real world. Namely, a physicist always begins by isolating the most important degrees of freedom to describe a given system at a given energy and/or size scale. These degrees of freedom must be compatible with the ones that govern objects at a finer level of description, and must define the degrees of freedom useful at any coarser level of description. However, it is neither useful, nor sensible, nor fruitful, nor doable to overjump different levels. Why bother to derive the structure of a living cell from the unified QCD and electroweak Lagrangian? There are at least seven levels in between: nucleons are built of quarks, nuclei of nucleons, atoms of nuclei and electrons, molecules of atoms, amino acids of molecules, proteins of amino acids, and we did not arrive at a cell yet. Well, we shall not embark here on the philosophy of science; in the following we concentrate on describing how mesons and nucleons are built of quarks, and nuclei of nucleons.