© 20xx Elsevier Ltd. All rights reserved.
Figure 9. Hedgehog Skyrmion with radial isovectors
Nomenclature
Key Points
Section Introduction briefly explains the origin of chiral symmetry and its breaking in QCD.
Section Large-N C N_{C} N C QCD} reflects on QCD with a large number of color degrees of freedom and motives the soliton picture.
Section Solitons in field theory reviews the concept of solitons in non-linear field theories and connects to the motivation from Section Large-$N_{C}$ QCD.
Section Chiral Lagrangian introduces the concept of effective chiral Lagrangians which is the point of departure for chiral solitons.
Section The Skyrme model discusses the Skyrme model for baryons. This section conveys the main information about the soliton picture. Improvements, in the sense of better reproduction of empirical data, of this model are (very) technical but do not alter the soliton picture for baryons substantially. An exception is the chiral quark soliton discussed in Section Quarks and Skyrmions
Section Extension to SU(3) discusses the extension to three flavors by including kaons in the effective Lagrangian. The main objective is the exploration of the hyperon spectrum.
Section Vector mesons reflects on including light vector mesons. Rather than going into much of the technical details, the major improvements over the Skyrme model are listed.
Section Quarks and Skyrmions explains how a chiral soliton can be constructed from a model for the quark flavor dynamics. Here it is important to consider both, the energy from quark levels bound by the chiral soliton and the energy from the distorted Dirac sea because these two components emerge at the same order in the semi-classical expansion. A particular novelty is the possibility to explore nucleon structure functions.
Section Further applications lists further applications of chiral solitons in particle physics in compact form. The reader is referred to the quoted literature for more details.
Section Conclusion briefly summarizes the article.
Introduction
Ordinary matter is composed of molecules formed from atoms. An atom in turn is described as an electron cloud bound to a nucleus through the electromagnetic interaction. The nucleus itself is built from protons and neutrons. The latter are members of a larger group of (formerly considered elementary) particles: the baryons. Together with the mesons (pions, kaons, etc.) the baryons constitute the even larger group of hadrons that comprise all particles which are subject to the strong nuclear force. The quark model [1 , 2 ] is a systematic description of hadrons. Quarks are spin 1 2 \frac{1}{2} 2 1 particles and three quarks comprise baryons, which are known as half-integer spin objects, while mesons (integer spin) are quark-antiquark compounds. There are different quark species, called flavor, to accommodate the various quantum numbers of the hadrons. In total there are six flavors, though in this context we are mainly concerned with the three light ones: up , down and strange , cf., however, Section Heavy flavor symmetry. Pauli’s exclusion principle for fermions dictates an anti-symmetric wave-function and, since the ground state baryons have a symmetric spatial wave-function, this requires an additional quantum number for the quarks, nowadays known as color. The fundamental field theory for the interaction of quarks is the non-Abelian gauge theory for this color degree of freedom [3 , 4 ] : quantum chromodynamics (QCD). Since three quarks make up the anti-symmetric wave-function, there must be equally many color degrees of freedom and therefore QCD is an S U ( 3 ) SU(3) S U ( 3 ) gauge theory. This theory is akin to quantum electrodynamics which governs the interaction of electrons and photons. The quarks are associated with the electrons and the photon equivalent is called gluon. The main difference, however, is that while photons are electrically neutral, the gluons carry color charge and there are eight of them. The quarks spinors, q f q_f q f and gluon fields G μ = ∑ a = 1 8 G μ a λ a 2 G_{\mu} = \sum_{a=1}^{8}G_{\mu}^{a}\frac{\lambda^{a}}{2} G μ = ∑ a = 1 8 G μ a 2 λ a are respectively combined in the fundamental and adjoint representations of S U ( 3 ) SU(3) S U ( 3 ) . While the label ”f f f ” stands for flavor, the Dirac and color indexes of the spinors are not written out. The matrix character of the gluon field is represented by the Gell-Mann matrices, λ a \lambda_a λ a . The indexes of these matrices are contracted with the color index of the spinors. Then the QCD Lagrangian reads
L Q C D = − 1 2 tr ( G μ ν G μ ν ) + ∑ f q ˉ f γ μ ( ∂ μ − i g G μ ) q f − m f q ˉ f q f + gauge fixing terms . (1) \mathcal{L}_{\mathrm{QCD}}=-\frac{1}{2}\operatorname{tr}\left(G_{\mu\nu}G^{\mu\nu}\right)+\sum_f\bar{q}_f\gamma^\mu\left(\partial_\mu-igG_\mu\right)q_f-m_f\bar{q}_fq_f+\textbf{gauge fixing terms}. \tag*{(1)} L QCD = − 2 1 tr ( G μν G μν ) + f ∑ q ˉ f γ μ ( ∂ μ − i g G μ ) q f − m f q ˉ f q f + gauge fixing terms . (1)
The matrix valued field strength tensor, G μ ν = ∂ μ G ν − ∂ ν G μ − i g [ G μ , G ν ] G_{\mu\nu}=\partial_\mu G_\nu-\partial_\nu G_\mu-ig\left[G_\mu,G_\nu\right] G μν = ∂ μ G ν − ∂ ν G μ − i g [ G μ , G ν ] induces three- and four-point gluon self-interactions with strengths g g g and g 2 g^2 g 2 , respectively. These interactions make the main difference to quantum electrodynamics. While in the large energy regime these self-interactions lead to asymptotic freedom [5 , 6 ] , they hamper the perturbative treatment at low energies, which is the standard approach to explore quantum field theories. This subtlety is reflected by the confinement hypothesis: Quarks and gluons do not exist as isolated states, rather they combine to hadrons. In a group theoretical language this translates into the statement that only color singlets are asymptotically observable states.
The above discussion indicates that QCD is not too helpful in gaining insight into (static) baryon properties like radii or electromagnetic moments1 and we should instead explore QCD inspired models for that purpose. Besides the obvious Lorentz-invariance and the above described color gauge symmetry, the Lagrangian, Eq. (1) embodies a further symmetry when m f = 0 m_f=0 m f = 0 which is called the chiral symmetry 2 . It becomes apparent when introducing left- and right handed components
Ψ L , R = 1 2 ( 1 ∓ γ 5 ) Ψ (2) \Psi_{L,R}=\frac{1}{2}\left(1\mp\gamma_5\right)\Psi \tag*{(2)} Ψ L , R = 2 1 ( 1 ∓ γ 5 ) Ψ (2)
of the Dirac spinors (for each flavor and color). Without the mass term and with only vector interactions, Ψ ˉ γ μ G μ Ψ \bar{\Psi}\gamma^\mu G_\mu\Psi Ψ ˉ γ μ G μ Ψ as in L Q C D \mathcal{L}_{\mathrm{QCD}} L QCD , these components decouple. The quark mass spectrum separates into two sets: (i) the light quarks (up and down current quark masses are few MeV, the strange quark mass parameter is about 100 MeV) and (ii) charm, bottom and top (∼ 1.3 G e V , 4.5 G e V , 170 G e V \sim1.3\,\mathrm{GeV}, 4.5\,\mathrm{GeV}, 170\,\mathrm{GeV} ∼ 1.3 GeV , 4.5 GeV , 170 GeV , respectively) [9 ] . Hence, at a typical energy scale of hadron interactions (several hundred MeV and more) chiral symmetry may well be respected by the up and down flavors and eventually also by the strange counterpart. In this picture the QCD Lagrangian possesses a U L ( N f ) × U R ( N f ) U_L(N_f)\times U_R(N_f) U L ( N f ) × U R ( N f ) symmetry. Here N f N_f N f is the number of quark flavors whose current quark masses may be ignored. Depending on whether we consider the strange current quark mass as large or small we have N f = 2 N_f=2 N f = 2 or N f = 3 N_f=3 N f = 3 . This symmetry group factorizes according to
U L ( N f ) × U R ( N f ) ≃ U L + R ( 1 ) × U L − R ( 1 ) × S U L ( N f ) × S U R ( N f ) (3) U_L(N_f)\times U_R(N_f)\simeq U_{L+R}(1)\times U_{L-R}(1)\times S U_L(N_f)\times S U_R(N_f) \tag*{(3)} U L ( N f ) × U R ( N f ) ≃ U L + R ( 1 ) × U L − R ( 1 ) × S U L ( N f ) × S U R ( N f ) (3)
and is called the chiral group . The invariance under U L + R ( 1 ) U_{L+R}(1) U L + R ( 1 ) is responsible for the conservation of baryon number whereas U L − R ( 1 ) U_{L-R}(1) U L − R ( 1 ) is subject to a quantum anomaly [10 , 11 ] . This results in 2 N f − 1 2N_f-1 2 N f − 1 conserved flavor currents. The 2 N f 2N_f 2 N f flavor currents are most conveniently presented as linear combination of the left- and right-handed vector currents that have definite parity: the vector current J a μ J^\mu_a J a μ and the axial vector current A a μ A^\mu_a A a μ
J a μ = q ˉ L γ μ λ a 2 q L + q ˉ R γ μ λ a 2 q R = q ˉ γ μ λ a 2 q and A a μ = − q ˉ L γ μ λ a 2 q L + q ˉ R γ μ λ a 2 q R = q ˉ γ μ γ 5 λ a 2 q . (4) J^\mu_a=\bar{q}_L\gamma^\mu\frac{\lambda_a}{2}q_L+\bar{q}_R\gamma^\mu\frac{\lambda_a}{2}q_R=\bar{q}\gamma^\mu\frac{\lambda_a}{2}q
\qquad\text{and}\qquad
A^\mu_a=-\bar{q}_L\gamma^\mu\frac{\lambda_a}{2}q_L+\bar{q}_R\gamma^\mu\frac{\lambda_a}{2}q_R=\bar{q}\gamma^\mu\gamma_5\frac{\lambda_a}{2}q. \tag*{(4)} J a μ = q ˉ L γ μ 2 λ a q L + q ˉ R γ μ 2 λ a q R = q ˉ γ μ 2 λ a q and A a μ = − q ˉ L γ μ 2 λ a q L + q ˉ R γ μ 2 λ a q R = q ˉ γ μ γ 5 2 λ a q . (4)
Here λ a \lambda_a λ a (a = 1 , … , N f 2 − 1 a=1,\ldots,N_f^2-1 a = 1 , … , N f 2 − 1 ) are the Gell-Mann matrices of S U ( N f ) S U(N_f) S U ( N f ) and λ 0 = 2 / N f 1 N f × N f \lambda_0=\sqrt{2/N_f}\,\mathbf{1}_{N_f\times N_f} λ 0 = 2/ N f 1 N f × N f is proportional to the unit matrix in flavor space. In the above construction the spinors are column vectors with N f N_f N f entries of the quark flavors
q = ( Ψ 1 Ψ 2 ⋮ Ψ N f ) = ( Ψ u Ψ d ⋮ Ψ N f ) . (5) q=\begin{pmatrix}\Psi_1\\\Psi_2\\\vdots\\\Psi_{N_f}\end{pmatrix}
=\begin{pmatrix}\Psi_u\\\Psi_d\\\vdots\\\Psi_{N_f}\end{pmatrix}. \tag*{(5)} q = Ψ 1 Ψ 2 ⋮ Ψ N f = Ψ u Ψ d ⋮ Ψ N f . (5)
In the construction of the currents, Eq. (4) the color label of the spinors is simply summed over. The vector current, J a μ J^\mu_a J a μ is conserved when the quark masses are equal while the conservation of the axial current, A a μ A^\mu_a A a μ emerges only when all masses are fully ignored, while ∂ μ A a μ ≠ 0 \partial_\mu A^\mu_a\ne0 ∂ μ A a μ = 0 due to the above mentioned anomaly.
The symmetry structure, Eq. (3) suggests that QCD predicts two sets of matter that, up to minor corrections from m f ≠ 0 m_f\ne0 m f = 0 , have the same mass spectrum. Hadrons in one set are composed of left-handed quarks, the other set has the right-handed partners. Hence, the elements of the two sets differ by their parity quantum number. This scenario would result from the Wigner-Weyl realization of chiral symmetry when the lowest energy configuration (vacuum) is invariant under the symmetry and the generators of the symmetry, i.e., the Noether charges constructed from the currents in Eq. (4) , transform degenerate physical states into one another. However, Nature is different: There are low-lying pseudoscalar mesons like the pions (m π ≈ 135 M e V m_\pi\approx135\,\mathrm{MeV} m π ≈ 135 MeV ) and the kaons (m K ≈ 495 M e V m_K\approx495\,\mathrm{MeV} m K ≈ 495 MeV ) but no scalar mesons with similar masses exist [9 ] . Hence we conclude that chiral symmetry is subject to the Nambu-Goldstone realization: the vacuum is not invariant under the symmetry, rather massless bosons (called Goldstone bosons) are excited. This scenario is called spontaneous chiral symmetry breaking.
Furthermore non-singlet operators possess non-zero VEVs. Vector transformations do not mix left- and right-handed spinors while axial transformations do. Observing that
Ψ ‾ Ψ = Ψ ‾ L Ψ R + Ψ ‾ R Ψ L , (6) \overline{\Psi}\Psi=\overline{\Psi}_{L}\Psi_{R}+\overline{\Psi}_{R}\Psi_{L}, \tag*{(6)} Ψ Ψ = Ψ L Ψ R + Ψ R Ψ L , (6)
it is perspicuous that the simplest non-singlet operator is q ‾ q \overline{q}q q q and that the dynamics of QCD implies a non-zero quark condensate,
⟨ q ‾ q ⟩ ≠ 0. (7) \langle\overline{q}q\rangle\ne0. \tag*{(7)} ⟨ q q ⟩ = 0. (7)
This non-zero VEV is called the quark condensate. Model building therefore requires to (i) find a simple mechanism that yields such a VEV or (ii) start from a formulation that has Eq. (7) built in.
Large-N C N_C N C QCD
The number of color degrees of freedom can be generalized from three to N C N_C N C so that 1 N C \frac{1}{N_C} N C 1 becomes a hidden expansion parameter for QCD [12 ] . This expansion is mainly a matter of combinatorics. The leading large-N C N_C N C components of the combinatoric factors are obtained by treating gluons as quark-antiquark compounds where each may have an arbitrary color quantum number. In Feynman diagrams (anti)quark lines maintain their color quantum number. Demanding that the gluon propagator has a smooth large-N C N_C N C limit, stipulates the scaling g ∼ 1 N C g\sim\frac{1}{\sqrt{N_C}} g ∼ N C 1 and leads to three important observations for the Feynman diagrams that are relevant in the large-N C N_C N C limit [13 ] : (i) internal lines are made from gluons, (ii) all lines are in a single plane, and (iii) the edges only have quark lines. These diagrams scale with N C N_C N C . Figure 1 (PDF p. 4) shows two typical higher order interaction diagrams that contribute to the gluon propagator. The left diagram is planar and O ( N C ) O(N_C) O ( N C ) , but the right diagram is non-planar and O ( 1 / N C ) O(1/N_C) O ( 1/ N C ) . The fact that for large-N C N_C N C neither non-planar nor diagrams with internal quark loops contribute
Figure 1 (PDF p. 4). A planar (left) and a non-planar (right) diagram that contribute to the Feynman diagrams for the gluon propagator. The dots indicate three- and four-point gluon self-interaction vertices.
causes the color indices along any Cutkosky-cut of a Feynman diagram for a correlation function to be combined in a single trace, rather than products of traces. This translates into the statement that in the large-N C N_C N C limit all intermediate states are quark bilinear color singlets objects, i.e., single mesons, taking for granted that QCD is a confining theory. The absence of multiple-meson intermediate states implies that the quadratic correlation function of a quark bilinear operator J ( x ) J(x) J ( x ) that is inserted into the quark line at the edge has the simple spectral decomposition
⟨ J ( x ) J ( y ) ⟩ = ∫ d 4 k ( 2 π ) 4 e i k ( x − y ) ∑ i a i ( k ) a i † ( k ) k 2 − m i 2 . (8) \langle J(x)J(y)\rangle=\int\frac{d^{4}k}{(2\pi)^{4}}e^{ik(x-y)}\sum_i\frac{a_i(k)a_i^\dagger(k)}{k^{2}-m_i^{2}}. \tag*{(8)} ⟨ J ( x ) J ( y )⟩ = ∫ ( 2 π ) 4 d 4 k e ik ( x − y ) i ∑ k 2 − m i 2 a i ( k ) a i † ( k ) . (8)
The above sum goes over all meson states that can couple to J ( x ) J(x) J ( x ) and a i ∼ ⟨ 0 ∣ J ∣ i ⟩ a_i\sim\langle0|J|i\rangle a i ∼ ⟨ 0∣ J ∣ i ⟩ is the amplitude for J J J to create the meson i i i (with mass m i m_i m i ) from the vacuum. The left-hand-side of Eq. (8) is linear in N C N_C N C so that a i ∼ N C a_i\sim\sqrt{N_C} a i ∼ N C and lim N C → ∞ m i = c o n s t \lim_{N_C\to\infty}m_i=\mathrm{const} lim N C → ∞ m i = const . On the other hand, infinitely many mesons must contribute in the sum, Eq. (8) , in order to reproduce the ln ( k 2 ) \ln(k^2) ln ( k 2 ) behavior in the asymptotically free regime of QCD.
A diagram with n n n insertions of J ( x ) J(x) J ( x ) scales with N C N_C N C as well and we can write
⟨ 0 ∣ J ∣ i 1 ⟩ … ⟨ 0 ∣ J ∣ i n ⟩ ⏟ n t e r m s Γ i 1 … i n ( n ) = O ( N C n 2 ) × Γ i 1 … i n ( n ) = O ( N C ) . (9) \underbrace{\langle0|J|i_1\rangle\ldots\langle0|J|i_n\rangle}_{n\ \mathrm{terms}}\Gamma^{(n)}_{i_1\ldots i_n}=O\left(N_C^{\frac{n}{2}}\right)\times\Gamma^{(n)}_{i_1\ldots i_n}=O(N_C). \tag*{(9)} n terms ⟨ 0∣ J ∣ i 1 ⟩ … ⟨ 0∣ J ∣ i n ⟩ Γ i 1 … i n ( n ) = O ( N C 2 n ) × Γ i 1 … i n ( n ) = O ( N C ) . (9)
Thus the coupling constant for a vertex with n n n mesons scales as
Γ i 1 … i n ( n ) = O ( N C 1 − n 2 ) . (10) \Gamma^{(n)}_{i_1\ldots i_n}=O\left(N_C^{1-\frac{n}{2}}\right). \tag*{(10)} Γ i 1 … i n ( n ) = O ( N C 1 − 2 n ) . (10)
In particular the four-point function vanishes like Γ ( 4 ) = g e f f ∼ 1 N C \Gamma^{(4)}=g_{\mathrm{eff}}\sim\frac{1}{N_C} Γ ( 4 ) = g eff ∼ N C 1 in the large N C N_C N C -limit. This leads to the conjecture [13 ] that QCD is equivalent to an effective weakly (i.e., g e f f → 0 g_{\mathrm{eff}}\to0 g eff → 0 ) interacting meson theory. The caveat of having infinitely many mesons should not be a problem in the low energy regime where only a few light mesons are involved.
The large-N C N_C N C description of baryons is more complicated because we need (at least) N C N_C N C quark constituents to form a color singlet. In turn the relevant Feynman diagrams like those in Figure 2 do not have a finite limit as N C N_C N C approaches infinity. More specifically we expect a diagram with m m m single gluon exchanges to scale as
Figure 2. Feynman diagram for a propagating baryon built from N C N_C N C quarks with one (left) and two (right) gluon exchanges. Interaction vertices are shown as big dots.
g 2 m m ! [ N C ( N C − 1 ) ] m = O ( N C m ) . (11) \frac{g^{2m}}{m!}\left[N_C(N_C-1)\right]^m = O\left(N_C^m\right). \tag*{(11)} m ! g 2 m [ N C ( N C − 1 ) ] m = O ( N C m ) . (11)
The proper way to sum all contributions uses many body techniques as explained in Ref. [13 ] . Since we are solely interested in the large-N C N_C N C combinatorics it suffices to consider the many body problem in a non-relativistic Hartree approach. In that framework a single quark reacts on an average potential generated by the N C − 1 N_C-1 N C − 1 remaining quarks. In the non-relativistic case only the two-body forces contribute and spin dependent forces may been omitted. Then the Hamilton operator reads
H = N C M + ∑ i = 1 N C − ∂ i 2 2 M − g s 2 N C ∑ i < j N C 1 ∣ r i − r j ∣ , (12) H=N_C M+\sum_{i=1}^{N_C}\frac{-\partial_i^2}{2M}-\frac{g_s^2}{N_C}\sum_{i<j}^{N_C}\frac{1}{\lvert r_i-r_j\rvert}, \tag*{(12)} H = N C M + i = 1 ∑ N C 2 M − ∂ i 2 − N C g s 2 i < j ∑ N C ∣ r i − r j ∣ 1 , (12)
where M M M is the mass of a single quark and g s = N C g g_s=\sqrt{N_C}g g s = N C g is O ( N C 0 ) O(N_C^0) O ( N C 0 ) . In the ground state all quarks should be in the S S S -wave channel. This motivates the ansatz for the (scalar) many body wave-function (anti-symmetrization for color labels not explicitly written):
Ψ ( r 1 , … , r N C ) = ∏ i = 1 N C ϕ ( r i ) . (13) \Psi(r_1,\ldots,r_{N_C})=\prod_{i=1}^{N_C}\phi(r_i). \tag*{(13)} Ψ ( r 1 , … , r N C ) = i = 1 ∏ N C ϕ ( r i ) . (13)
We write E = N C ϵ E=N_C\epsilon E = N C ϵ and attempt to apply the variational principle to
⟨ Ψ ∣ H − E ∣ Ψ ⟩ = − N C ϵ + N C M + N C 2 M ∫ d 3 r ∂ ϕ ∗ ( r ) ⋅ ∂ ϕ ( r ) − N C ( N C − 1 ) 2 g s 2 N C ∫ d 3 r 1 ∫ d 3 r 2 ∣ ϕ ( r 1 ) ∣ 2 ∣ ϕ ( r 2 ) ∣ 2 ∣ r 1 − r 2 ∣ . (14) \langle\Psi\lvert H-E\rvert\Psi\rangle=-N_C\epsilon+N_C M+\frac{N_C}{2M}\int d^3r\,\partial\phi^*(r)\cdot\partial\phi(r)-\frac{N_C(N_C-1)}{2}\frac{g_s^2}{N_C}\int d^3r_1\int d^3r_2\,\frac{\lvert\phi(r_1)\rvert^2\lvert\phi(r_2)\rvert^2}{\lvert r_1-r_2\rvert}. \tag*{(14)} ⟨ Ψ ∣ H − E ∣ Ψ ⟩ = − N C ϵ + N C M + 2 M N C ∫ d 3 r ∂ ϕ ∗ ( r ) ⋅ ∂ ϕ ( r ) − 2 N C ( N C − 1 ) N C g s 2 ∫ d 3 r 1 ∫ d 3 r 2 ∣ r 1 − r 2 ∣ ∣ ϕ ( r 1 ) ∣ 2 ∣ ϕ ( r 2 ) ∣ 2 . (14)
As N C ≫ 1 N_C\gg1 N C ≫ 1 we pull out an overall factor N C N_C N C and the remaining variational principle then has a solution with both ϕ \phi ϕ and ϵ \epsilon ϵ being O ( N C 0 ) O(N_C^0) O ( N C 0 ) . This implies that baryon masses are O ( N C ) O(N_C) O ( N C ) . The fact that ϕ \phi ϕ has a smooth large-N C N_C N C limit causes the typical extension of a baryon ⟨ r 2 ⟩ = ⟨ Ψ ∣ ∑ i r i 2 ∣ Ψ ⟩ / N C \langle r^2\rangle=\langle\Psi\lvert\sum_i r_i^2\lvert\Psi\rangle/N_C ⟨ r 2 ⟩ = ⟨ Ψ ∣ ∑ i r i 2 ∣ Ψ ⟩ / N C to be O ( N C 0 ) O(N_C^0) O ( N C 0 ) . The gluon self-interactions may indeed give rise to three and four body forces. In leading order the respective combinatoric factors are N C 3 N_C^3 N C 3 and N C 4 N_C^4 N C 4 . The coupling constants in the Hartree Hamiltonian are g 4 2 = g s 4 / N C 2 g_4^2=g_s^4/N_C^2 g 4 2 = g s 4 / N C 2 and g 6 2 = g s 6 / N C 3 g_6^2=g_s^6/N_C^3 g 6 2 = g s 6 / N C 3 . Hence these interactions add terms O ( N C ) O(N_C) O ( N C ) to Eq. (14) but do not lead to higher powers of N C N_C N C . In a similar fashion, Ref. [13 ] shows that the baryon-baryon scattering amplitude is also O ( N C ) O(N_C) O ( N C ) . The situation is different for meson-baryon scattering. Since we can only pick a single quark from the meson, the one gluon exchange contribution to the energy functional is O ( N C 0 ) O(N_C^0) O ( N C 0 ) , i.e., of the same order as the meson mass but suppressed compared to the baryon mass. Hence only the meson reacts in the scattering process while the (infinitely heavy) baryon is essentially unaffected. Thus the baryon piece ϕ \phi ϕ in the N C + 2 N_C+2 N C + 2 body Hartree wave-function
Ψ ( r 1 , … , r N C ; x , y , t ) = ∑ P ( − 1 ) P [ ∏ i = 1 N C ϕ ( r i ) ] u ( x , y , t ) (15) \Psi(r_1,\ldots,r_{N_C};x,y,t)=\sum_P(-1)^P\left[\prod_{i=1}^{N_C}\phi(r_i)\right]u(x,y,t) \tag*{(15)} Ψ ( r 1 , … , r N C ; x , y , t ) = P ∑ ( − 1 ) P [ i = 1 ∏ N C ϕ ( r i ) ] u ( x , y , t ) (15)
is the same as in Eq. (13) up to corrections of O ( 1 / N C ) O(1/N_C) O ( 1/ N C ) . The anti-symmetrization is with respect to the color labels of the quarks in the baryon and the quark in the meson. Then the wave-function is symmetric under the exchange of the quarks’ spatial coordinates, r i r_i r i and x x x . The meson part of the wave-function u ( x , y , t ) u(x,y,t) u ( x , y , t ) satisfies a linear integro-differential equation [13 ] ,
i ∂ ∂ t u ( x , y , t ) = − 1 2 M ( ∂ x 2 + ∂ y 2 ) u ( x , y , t ) − g s 2 u ( x , y , t ) ∣ x − y ∣ − g s 2 ϕ ( x ) ∫ d 3 z ϕ ∗ ( z ) u ( x , y , t ) [ 1 ∣ z − x ∣ + 1 ∣ z − y ∣ ] . (16) i\frac{\partial}{\partial t}u(x,y,t)=-\frac{1}{2M}\left(\partial_x^2+\partial_y^2\right)u(x,y,t)-g_s^2\frac{u(x,y,t)}{\lvert x-y\rvert}-g_s^2\phi(x)\int d^3z\,\phi^*(z)u(x,y,t)\left[\frac{1}{\lvert z-x\rvert}+\frac{1}{\lvert z-y\rvert}\right]. \tag*{(16)} i ∂ t ∂ u ( x , y , t ) = − 2 M 1 ( ∂ x 2 + ∂ y 2 ) u ( x , y , t ) − g s 2 ∣ x − y ∣ u ( x , y , t ) − g s 2 ϕ ( x ) ∫ d 3 z ϕ ∗ ( z ) u ( x , y , t ) [ ∣ z − x ∣ 1 + ∣ z − y ∣ 1 ] . (16)
The non-symmetric appearance of the coordinates x x x and y y y under the integral stems from the fact that the anti-quark coordinate y y y is not subject to anti-symmetrization in Eq. (15). Eq. (16) describes the scattering of a meson in a background potential parameterized by ϕ ( r ) \phi(r) ϕ ( r ) , i.e., generated by the baryon field. Obviously u ( x , y , t ) u(x,y,t) u ( x , y , t ) is O ( N C 0 ) O(N_C^0) O ( N C 0 ) and so are the meson-baryon scattering data that are extracted thereof. Combining these results with the large-N C N_C N C scaling behavior of baryon properties with the discussion about QCD is equivalent to an effective meson theory with coupling constant g e f f ∼ 1 N C g_{\mathrm{eff}}\sim\frac{1}{N_C} g eff ∼ N C 1 , we find that baryon masses are O ( 1 / g e f f ) O(1/g_{\mathrm{eff}}) O ( 1/ g eff ) while radii and meson-baryon scattering amplitudes approach constants at large-N C N_C N C .
Solitons in field theory
In this section we will exemplify that soliton solutions in meson theories match the above derived coupling constant dependences for baryon properties. For this purpose we briefly review the so-called kink as a simple example for a classical soliton in D = 1 + 1 D=1+1 D = 1 + 1 dimensions. Ref. [14 ] provides a more thorough discussion and additional examples. The model Lagrangian contains a fourth order self-interaction
L k i n k = 1 2 [ ( ∂ t Φ ) 2 − ( ∂ x Φ ) 2 ] − λ 4 ( Φ 2 − m 2 λ ) 2 (17) \mathcal{L}_{\mathrm{kink}}=\frac{1}{2}\left[(\partial_t\Phi)^2-(\partial_x\Phi)^2\right]-\frac{\lambda}{4}\left(\Phi^2-\frac{m^2}{\lambda}\right)^2 \tag*{(17)} L kink = 2 1 [ ( ∂ t Φ ) 2 − ( ∂ x Φ ) 2 ] − 4 λ ( Φ 2 − λ m 2 ) 2 (17)
for the scalar field Φ \Phi Φ . Obviously we identify λ ∼ g e f f \lambda\sim g_{\mathrm{eff}} λ ∼ g eff as the effective meson coupling constant, Γ ( 4 ) \Gamma^{(4)} Γ ( 4 ) for two particle scattering. To complete the analogy with the large-N C N_C N C discussion we demand that the mass parameter m m m remains constant when λ \lambda λ is changed.
There are two distinct vacuum configurations Φ = ± m λ \Phi=\pm\frac{m}{\sqrt{\lambda}} Φ = ± λ m . It is straightforward to verify that the Lagrangian, Eq. (17) has static solutions
Φ ± ( x ) = ± m λ tanh ( m 2 x ) (18) \Phi_{\pm}(x)=\pm\frac{m}{\sqrt{\lambda}}\tanh\left(\frac{m}{\sqrt{2}}x\right) \tag*{(18)} Φ ± ( x ) = ± λ m tanh ( 2 m x ) (18)
that mediate between these vacua. The profile function with the positive slope is called the kink soliton, the other antikink. The energy density (equal to the negative Lagrange density for static fields) is ϵ ( x ) = m 4 2 λ sech 4 ( m 2 x ) \epsilon(x)=\frac{m^4}{2\lambda}\operatorname{sech}^4\left(\frac{m}{\sqrt{2}}x\right) ϵ ( x ) = 2 λ m 4 sech 4 ( 2 m x ) . The classical energy and the average extension are computed as spatial integrals involving this density:
E c l = ∫ − ∞ ∞ d x ϵ ( x ) = 2 2 3 m 3 λ and ⟨ x 2 ⟩ = 1 E c l ∫ − ∞ ∞ d x x 2 ϵ ( x ) = 1 m 2 ( π 2 6 − 1 ) (19) E_{\mathrm{cl}}=\int_{-\infty}^{\infty}dx\,\epsilon(x)=\frac{2\sqrt{2}}{3}\frac{m^3}{\lambda}
\qquad\text{and}\qquad
\langle x^2\rangle=\frac{1}{E_{\mathrm{cl}}}\int_{-\infty}^{\infty}dx\,x^2\epsilon(x)=\frac{1}{m^2}\left(\frac{\pi^2}{6}-1\right) \tag*{(19)} E cl = ∫ − ∞ ∞ d x ϵ ( x ) = 3 2 2 λ m 3 and ⟨ x 2 ⟩ = E cl 1 ∫ − ∞ ∞ d x x 2 ϵ ( x ) = m 2 1 ( 6 π 2 − 1 ) (19)
Scattering is described by expanding around the kink, Φ ( x , t ) = Φ + ( x ) + η ( x , t ) \Phi(x,t)=\Phi_+(x)+\eta(x,t) Φ ( x , t ) = Φ + ( x ) + η ( x , t ) to harmonic order so that the field equation for the fluctuation η ( x , t ) \eta(x,t) η ( x , t ) reads
[ ∂ t 2 − ∂ x 2 + m 2 + 3 m 2 tanh 2 ( m 2 x ) ] η ( x , t ) = 0. (20) \left[\partial_t^2-\partial_x^2+m^2+3m^2\tanh^2\left(\frac{m}{\sqrt{2}}x\right)\right]\eta(x,t)=0 . \tag*{(20)} [ ∂ t 2 − ∂ x 2 + m 2 + 3 m 2 tanh 2 ( 2 m x ) ] η ( x , t ) = 0. (20)
This wave-equation does not contain λ \lambda λ and thus the scattering data extracted from η ( x , t ) \eta(x,t) η ( x , t ) are O ( λ 0 ) \mathcal{O}(\lambda^0) O ( λ 0 ) .
When we identify the soliton as a baryon with mass E c l E_{\mathrm{cl}} E cl and its fluctuations as scattering mesons, we recognize that this construction matches all the results from large-N C N_C N C QCD. The program is now set out: construct a chirally invariant theory for the low-mass mesons and find its soliton solutions.
Chiral Lagrangian
In this section we will sketch the construction of a chirally invariant Lagrangian for the low mass pions that would be Goldstone bosons in the limit of m u = m d = 0 m_u=m_d=0 m u = m d = 0 . Nevertheless we start from non-zero quark masses and write the mass term using Eq. (6) as
∑ f m f q ‾ f q f = ∑ i , j = 1 2 Ψ ‾ L , i M i j Ψ R , i + Ψ ‾ R , i M j i ∗ Ψ L , j with M = diag ( m u , m d ) (21) \sum_f m_f\overline{q}_f q_f=\sum_{i,j=1}^{2}\overline{\Psi}_{L,i}\mathcal{M}_{ij}\Psi_{R,i}+\overline{\Psi}_{R,i}\mathcal{M}^{*}_{ji}\Psi_{L,j}
\qquad\text{with}\qquad
\mathcal{M}=\operatorname{diag}(m_u,m_d) \tag*{(21)} f ∑ m f q f q f = i , j = 1 ∑ 2 Ψ L , i M ij Ψ R , i + Ψ R , i M j i ∗ Ψ L , j with M = diag ( m u , m d ) (21)
The lower-case Roman letters are the flavor labels for the up and down quarks (1 → u 1\to u 1 → u , 2 → d 2\to d 2 → d ), cf. Eq. (5) . Eventually we will also include strange flavors with M = diag ( m u , m d , m s ) \mathcal{M}=\operatorname{diag}(m_u,m_d,m_s) M = diag ( m u , m d , m s ) . With these expressions for the mass matrix M \mathcal{M} M , its conjugation in Eq. (21) is redundant. But when we consider M \mathcal{M} M as a more general quantity (called the spurion field) that transforms as M → L M R † \mathcal{M}\to L\mathcal{M}R^\dagger M → L M R † , where L L L and R R R are constant unitary matrices that respectively mix left- and right-handed fermion flavors, the mass term is no longer chirally invariant. We use the same language to write the transformed quark condensate, Eq. (7) as
Ψ ‾ R , i Ψ L , i = − σ δ i j → − σ ( L R † ) i j = − σ U i j (22) \overline{\Psi}_{R,i}\Psi_{L,i}=-\sigma\delta_{ij}\to-\sigma(LR^\dagger)_{ij}=-\sigma U_{ij} \tag*{(22)} Ψ R , i Ψ L , i = − σ δ ij → − σ ( L R † ) ij = − σ U ij (22)
where σ \sigma σ is a (positive) constant with cubic mass dimension. Note that U U U is also unitary. From Eq. (4) we observe that left- and right-handed fermions transform in the same way for vector transformations, i.e., for L = R L=R L = R and the condensate remains unchanged. In case L ≠ R L\ne R L = R the transformed condensate describes a different vacuum that would be degenerate for M = 0 \mathcal{M}=0 M = 0 . Applying space-time dependent axial transformations to this VEV generates the fields of the Goldstone bosons ϕ a ( x ) \phi_a(x) ϕ a ( x ) . Hence we should introduce these bosons via
U ( x ) = exp [ i ϕ a ( x ) τ a / f π ] , where the τ a are the Pauli matrices. (23) U(x)=\exp\left[i\phi_a(x)\tau_a/f_\pi\right],
\qquad\text{where the $\tau_a$ are the Pauli matrices.} \tag*{(23)} U ( x ) = exp [ i ϕ a ( x ) τ a / f π ] , where the τ a are the Pauli matrices. (23)
The constant f π f_\pi f π will be linked to an observable shortly. The next step is to establish a Lagrangian for the Goldstone boson fields ϕ a ( x ) \phi_a(x) ϕ a ( x ) . From the above discussions we deduce that under constant chiral rotations the so-called chiral field transforms as U → L U R † U\to LUR^\dagger U → LU R † . For M = 0 \mathcal{M}=0 M = 0 , that should be an invariance, and since U U † = 1 2 × 2 UU^\dagger=1_{2\times2} U U † = 1 2 × 2 , there cannot be a potential term in that case. From Lorentz symmetry the leading term must have two derivatives and the candidates are
[ tr ( U ∂ μ U † ) ] 2 , tr ( ∂ μ U ∂ μ U † ) and tr ( U ∂ μ U † U ∂ μ U † ) . \left[\operatorname{tr}\left(U\partial_\mu U^\dagger\right)\right]^2,\qquad\operatorname{tr}\left(\partial_\mu U\partial^\mu U^\dagger\right)\quad\text{and}\quad\operatorname{tr}\left(U\partial_\mu U^\dagger U\partial^\mu U^\dagger\right). [ tr ( U ∂ μ U † ) ] 2 , tr ( ∂ μ U ∂ μ U † ) and tr ( U ∂ μ U † U ∂ μ U † ) .
The first term is actually zero and the two others are proportional to each other because U ∂ μ U † = − ∂ μ U U † U\partial^\mu U^\dagger=-\partial^\mu U U^\dagger U ∂ μ U † = − ∂ μ U U † . For M ∝ 1 2 × 2 \mathcal{M}\propto\mathbb{1}_{2\times2} M ∝ 1 2 × 2 the vector transformation should still be an invariance and we can have terms like tr U \operatorname{tr}U tr U . Demanding the canonical normalization of the fields which we want to identify with the physical pions, the first guess effective meson Lagrangian reads
L σ = f π 2 4 tr ( ∂ μ U ∂ μ U † ) + f π 2 m π 2 4 tr ( U + U † − 21 2 × 2 ) . (24) \mathcal{L}_{\sigma}=\frac{f_\pi^2}{4}\operatorname{tr}\left(\partial_\mu U\partial^\mu U^\dagger\right)+\frac{f_\pi^2m_\pi^2}{4}\operatorname{tr}\left(U+U^\dagger-2\mathbb{1}_{2\times2}\right). \tag*{(24)} L σ = 4 f π 2 tr ( ∂ μ U ∂ μ U † ) + 4 f π 2 m π 2 tr ( U + U † − 2 1 2 × 2 ) . (24)
In the last term we have subtracted a constant so that the energy density of the to-be-constructed soliton vanishes at spatial infinity. Expanding this Lagrangian yields terms which are quartic in ϕ a \phi_a ϕ a or ∂ μ ϕ a \partial_\mu\phi_a ∂ μ ϕ a with coefficients proportional to m π 2 f π 2 \frac{m_\pi^2}{f_\pi^2} f π 2 m π 2 or 1 f π 2 \frac{1}{f_\pi^2} f π 2 1 . The consideration in the first part of Section Large NC QCD implies that f π = O ( N c ) f_\pi=O\left(\sqrt{N_c}\right) f π = O ( N c ) .
To get a number for f π f_\pi f π we use empirical information from the decay of the charged pions into muons (μ \mu μ ) and their (anti)neutrinos (ν \nu ν ). The relevant electroweak interaction Lagrangian at energies way below the W W W -mass is of a current-current form
L π → μ ν μ = G F 2 L α ( h a d r ) L ( l e p t ) , α , (25) \mathcal{L}_{\pi\to\mu\nu_\mu}=\frac{G_F}{\sqrt{2}}L_\alpha^{(\mathrm{hadr})}L^{(\mathrm{lept}),\alpha}, \tag*{(25)} L π → μ ν μ = 2 G F L α ( hadr ) L ( lept ) , α , (25)
where G F G_F G F is the Fermi constant and L α L_\alpha L α are the left-handed currents. The leptonic matrix element ⟨ 0 ∣ L ( l e p t ) , α ∣ μ ν ⟩ \langle0|L^{(\mathrm{lept}),\alpha}|\mu\nu\rangle ⟨ 0∣ L ( lept ) , α ∣ μν ⟩ is computed straightforwardly using Feynman rules. We get its hadronic counterpart from the currents in Eq. (4) as L α ( h a d r ) = J α a − A α a L_\alpha^{(\mathrm{hadr})}=J_\alpha^a-A_\alpha^a L α ( hadr ) = J α a − A α a , where the superscript denotes the pion charge a = 1 ± i 2 a=1\pm i2 a = 1 ± i 2 . By symmetry we have ⟨ 0 ∣ J a ( x ) ∣ π b ( p ) ⟩ = 0 \langle0|J^a(x)|\pi^b(p)\rangle=0 ⟨ 0∣ J a ( x ) ∣ π b ( p )⟩ = 0 , while the axial current matrix element defines the pion decay constant3 f π f_\pi f π via
⟨ 0 ∣ A μ a ( x ) ∣ π b ( q ) ⟩ = i f π δ a b q μ e − i q x . \langle0|A_\mu^a(x)|\pi^b(q)\rangle=if_\pi\delta_{ab}q_\mu e^{-iqx}. ⟨ 0∣ A μ a ( x ) ∣ π b ( q )⟩ = i f π δ ab q μ e − i q x .
Hence one can extract this constant from the pion decay width as f π ≈ 92.2 M e V f_\pi\approx92.2\ \mathrm{MeV} f π ≈ 92.2 MeV [9 ] .4 It remains to verify that this constant is indeed the one in the Lagrangian, Eq. (24) . To do so we need to get the axial current as the (would-be) Noether current of the transformation U → U + δ U U\to U+\delta U U → U + δ U with δ U = i 2 { τ a , U } ϵ a \delta U=\frac{i}{2}\{\tau_a,U\}\epsilon_a δ U = 2 i { τ a , U } ϵ a . The result is
A μ a ( x ) = f π 2 2 tr [ τ a 2 ( U † ∂ μ U − U ∂ μ U † ) ] = − f π ∂ μ ϕ a ( x ) + O ( ϕ 2 ) . (26) A_\mu^a(x)=\frac{f_\pi^2}{2}\operatorname{tr}\left[\frac{\tau^a}{2}\left(U^\dagger\partial_\mu U-U\partial_\mu U^\dagger\right)\right]=-f_\pi\partial_\mu\phi^a(x)+O\left(\phi^2\right). \tag*{(26)} A μ a ( x ) = 2 f π 2 tr [ 2 τ a ( U † ∂ μ U − U ∂ μ U † ) ] = − f π ∂ μ ϕ a ( x ) + O ( ϕ 2 ) . (26)
Taking its matrix element between the vacuum and a one-pion state with momentum q μ q_\mu q μ confirms the stated equality.
Here we used that we can relate the symmetry transformations of QCD to those in the effective theory. This allows us to compute relevant matrix elements of hadrons, even though we are not able to relate the hadron fields to the quark and gluon fields of QCD. This principal concept will recur when computing baryon properties.
The Skyrme model
Before introducing the Skyrme model and discussing (some of) its predictions for baryons we note that there are numerous review articles [15 , 16 , 17 , 18 , 19 ] and textbooks [20 , 21 , 22 ] that provide detailed background information and references to the original research publications.
The soliton
The next step is to find a soliton solution from the chiral Lagrangian, Eq. (24) . However, Derrick’s theorem [23 ] quickly shows that there is no such static solution. Assume that U 1 ( r ) U_1(r) U 1 ( r ) would be such as solution. Then the energy of U s ( r ) = U 1 ( s r ) U_s(r)=U_1(sr) U s ( r ) = U 1 ( sr ) must be minimal for s = 1 s=1 s = 1 . Straightforward calculation yields
E [ U s ] = f π 2 4 s ∫ d 3 r tr ( ∂ U 1 ) ⋅ ( ∂ U 1 † ) + f π 2 m π 2 4 s 3 ∫ d 3 r tr ( 21 2 × 2 − U 1 − U 1 † ) . (27) E[U_s]=\frac{f_\pi^2}{4s}\int d^3r\operatorname{tr}\left(\partial U_1\right)\cdot\left(\partial U_1^\dagger\right)+\frac{f_\pi^2m_\pi^2}{4s^3}\int d^3r\operatorname{tr}\left(2\mathbb{1}_{2\times2}-U_1-U_1^\dagger\right). \tag*{(27)} E [ U s ] = 4 s f π 2 ∫ d 3 r tr ( ∂ U 1 ) ⋅ ( ∂ U 1 † ) + 4 s 3 f π 2 m π 2 ∫ d 3 r tr ( 2 1 2 × 2 − U 1 − U 1 † ) . (27)
The two integrals are non-negative and therefore ∂ E [ U s ] ∂ s ≠ 0 \frac{\partial E[U_s]}{\partial s}\ne0 ∂ s ∂ E [ U s ] = 0 for any value of s s s . To remedy the problem, Skyrme added a chirally invariant term with four derivatives, but only two time derivatives [24 , 25 ] . Introducing α μ = U † ∂ μ U \alpha_\mu=U^\dagger\partial_\mu U α μ = U † ∂ μ U the Skyrme model Lagrangian reads
L S k = − f π 2 4 tr ( α μ α μ ) + 1 32 e 2 tr ( [ α μ , α ν ] [ α μ , α ν ] ) + f π 2 m π 2 4 tr ( U + U † − 21 2 × 2 ) . (28) \mathcal{L}_{\mathrm{Sk}}=-\frac{f_\pi^2}{4}\operatorname{tr}\left(\alpha_\mu\alpha^\mu\right)+\frac{1}{32e^2}\operatorname{tr}\left([\alpha_\mu,\alpha_\nu][\alpha^\mu,\alpha^\nu]\right)+\frac{f_\pi^2m_\pi^2}{4}\operatorname{tr}\left(U+U^\dagger-2\mathbb{1}_{2\times2}\right). \tag*{(28)} L Sk = − 4 f π 2 tr ( α μ α μ ) + 32 e 2 1 tr ( [ α μ , α ν ] [ α μ , α ν ] ) + 4 f π 2 m π 2 tr ( U + U † − 2 1 2 × 2 ) . (28)
The expansion of the additional, so-called Skyrme term in powers of ϕ a \phi_a ϕ a has the leading contribution proportional to 1 e 2 f π 4 ( ∂ ϕ a ) 4 \frac{1}{e^2 f_\pi^4}(\partial\phi_a)^4 e 2 f π 4 1 ( ∂ ϕ a ) 4 . Hence the Skyrme parameter scales as e = O ( 1 / N C ) e=O(1/\sqrt{N_C}) e = O ( 1/ N C ) .
To construct the soliton solution we consider a spherically symmetric ansatz which is called hedgehog configuration and was already explored by Pauli [26 ]
U ( x ) = U H ( x ) ≡ exp ( i τ ⃗ ⋅ x ^ F ( r ) ) with r = ∣ x ∣ . (29) U(x)=U_{\mathrm{H}}(x)\equiv\exp(i\vec{\tau}\cdot\hat{x}F(r)) \qquad\text{with} \qquad r=|x|. \tag*{(29)} U ( x ) = U H ( x ) ≡ exp ( i τ ⋅ x ^ F ( r )) with r = ∣ x ∣. (29)
The radial function F ( r ) F(r) F ( r ) is called the chiral angle. The chiral field is in flavor space, i.e., isospin for two flavors. The expression hedgehog refers to isovectors pointing away from the origin, similar to the spines of a hedgehog. This is indicated by the picture on the front page of this article. Substituting the ansatz from Eq. (29) yields the classical energy functional
E c l [ F ] = 2 π f π e ∫ 0 ∞ d x { x 2 F ′ 2 + 2 sin 2 F + 2 μ π 2 x 2 ( 1 − cos F ) + sin 2 F ( 2 F ′ 2 + sin 2 F x 2 ) } . (30) E_{\mathrm{cl}}[F]=\frac{2\pi f_\pi}{e}\int_0^\infty dx\left\{x^2F'^2+2\sin^2F+2\mu_\pi^2x^2(1-\cos F)+\sin^2F\left(2F'^2+\frac{\sin^2F}{x^2}\right)\right\}. \tag*{(30)} E cl [ F ] = e 2 π f π ∫ 0 ∞ d x { x 2 F ′2 + 2 sin 2 F + 2 μ π 2 x 2 ( 1 − cos F ) + sin 2 F ( 2 F ′2 + x 2 sin 2 F ) } . (30)
Dimensionless quantities x = e f π r x=ef_\pi r x = e f π r and μ π = m π / ( e f π ) \mu_\pi=m_\pi/(ef_\pi) μ π = m π / ( e f π ) have been introduced and the prime denotes the derivative with respect to x x x . Since μ π = O ( N C 0 ) \mu_\pi=O(N_C^0) μ π = O ( N C 0 ) , the N C N_C N C scaling is fully determined by the coefficient of the spatial integral and we immediately deduce that indeed E c l = O ( N C ) E_{\mathrm{cl}}=O(N_C) E cl = O ( N C ) . Upon variation we find the stationary condition,
( x 2 + 2 sin 2 F ) F ′ ′ + 2 x F ′ − sin 2 F − sin 2 F ( sin 2 F x 2 + F ′ 2 ) = μ π 2 x 2 sin F . (31) \left(x^2+2\sin^2F\right)F''+2xF'-\sin2F-\sin2F\left(\frac{\sin^2F}{x^2}+F'^2\right)=\mu_\pi^2x^2\sin F. \tag*{(31)} ( x 2 + 2 sin 2 F ) F ′′ + 2 x F ′ − sin 2 F − sin 2 F ( x 2 sin 2 F + F ′2 ) = μ π 2 x 2 sin F . (31)
Asymptotically the chiral field should approach the vacuum value U → 1 2 × 2 U\to\mathbb{1}_{2\times2} U → 1 2 × 2 . Linearizing the stationary condition leads to
F ( x ) → r → ∞ c ( 1 + μ π x ) e − μ π x x 2 , (32) F(x)\xrightarrow{r\to\infty}c(1+\mu_\pi x)\frac{e^{-\mu_\pi x}}{x^2}, \tag*{(32)} F ( x ) r → ∞ c ( 1 + μ π x ) x 2 e − μ π x , (32)
with some constant c c c . Though the pion mass term originates from the small explicit breaking of chiral symmetry and typically contributes only marginally to E c l E_{\mathrm{cl}} E cl , it induces an exponential decay of the chiral angle without which some of the integrals to be computed later would not converge.
We have determined the boundary condition for x → ∞ x\to\infty x → ∞ but still need to find F ( 0 ) F(0) F ( 0 ) . To this end we note that the Lagrangian, Eq. (28) does not reflect the pseudoscalar nature of the pion which relates to the simultaneous transformation of the fields and the spatial coordinates
ϕ a ( x , t ) → − ϕ a ( − x , t ) ⟺ U ( x , t ) → U † ( − x , t ) . \phi_a(x,t)\to-\phi_a(-x,t)\qquad\Longleftrightarrow\qquad U(x,t)\to U^\dagger(-x,t). ϕ a ( x , t ) → − ϕ a ( − x , t ) ⟺ U ( x , t ) → U † ( − x , t ) .
However, the action associated with the Lagrangian, Eq. (28) is invariant under x → − x x\to-x x → − x and U ( x , t ) → U † ( x , t ) U(x,t)\to U^\dagger(x,t) U ( x , t ) → U † ( x , t ) individually. To break that, a contribution with odd powers in α μ \alpha_\mu α μ would be needed, which is not possible in four space-time dimensions. On the level of the field equation the pseudoscalar nature can however be implemented. This is achieved by the additional λ \lambda λ term in
f π 2 2 ∂ μ ϕ a ∂ μ ϕ a + 1 8 e 2 ∂ μ ( α ν [ α μ , α ν ] ) − f π 2 m π 2 2 ( U − U † ) + 5 i λ ϵ μ ν ρ σ α μ α ν α ρ α σ = 0. (33) \frac{f_\pi^2}{2}\partial_\mu\phi^a\partial^\mu\phi^a+\frac{1}{8e^2}\partial_\mu\left(\alpha_\nu[\alpha^\mu,\alpha^\nu]\right)-\frac{f_\pi^2m_\pi^2}{2}(U-U^\dagger)+5i\lambda\epsilon_{\mu\nu\rho\sigma}\alpha^\mu\alpha^\nu\alpha^\rho\alpha^\sigma=0. \tag*{(33)} 2 f π 2 ∂ μ ϕ a ∂ μ ϕ a + 8 e 2 1 ∂ μ ( α ν [ α μ , α ν ] ) − 2 f π 2 m π 2 ( U − U † ) + 5 iλ ϵ μν ρ σ α μ α ν α ρ α σ = 0. (33)
The additional term does not effect the field equations in the static case. Even though it cannot be derived from the variation of a local Lagrangian, Witten showed in Refs. [27 , 28 ] that it can be obtained when applying the variational principle to the so-called Wess-Zumino term
Γ W Z = i λ ϵ μ ν ρ σ τ ∫ M 5 d 5 x tr [ α μ α ν α ρ α σ α τ ] . (34) \Gamma_{\mathrm{WZ}}=i\lambda\epsilon_{\mu\nu\rho\sigma\tau}\int_{M_5}d^5x\operatorname{tr}\left[\alpha^\mu\alpha^\nu\alpha^\rho\alpha^\sigma\alpha^\tau\right]. \tag*{(34)} Γ WZ = iλ ϵ μν ρ σ τ ∫ M 5 d 5 x tr [ α μ α ν α ρ α σ α τ ] . (34)
where the boundary of the five-dimensional manifold, M 5 M_5 M 5 is Minkowski space. The physics requirement that gauging Γ W Z \Gamma_{\mathrm{WZ}} Γ WZ with respect to the electromagnetic interaction reproduces the QCD result for the neutral pion decay π 0 → γ γ \pi^0\to\gamma\gamma π 0 → γ γ determines the so-far unknown coefficient5
Γ W Z = i N C 240 π 2 ϵ μ ν ρ σ τ ∫ M 5 d 5 x tr [ α μ α ν α ρ α σ α τ ] . (35) \Gamma_{\mathrm{WZ}}=i\frac{N_C}{240\pi^2}\epsilon_{\mu\nu\rho\sigma\tau}\int_{M_5}d^5x\operatorname{tr}\left[\alpha^\mu\alpha^\nu\alpha^\rho\alpha^\sigma\alpha^\tau\right]. \tag*{(35)} Γ WZ = i 240 π 2 N C ϵ μν ρ σ τ ∫ M 5 d 5 x tr [ α μ α ν α ρ α σ α τ ] . (35)
We recognize that each term in Eq. (33) is linear in N C N_C N C . As in the electromagnetic case, the interaction with a (background) baryon gauge field b μ b^\mu b μ is incorporated by the gauge principle and we need to gauge the U V ( 1 ) U_V(1) U V ( 1 ) symmetry. Since U U U by itself is not altered by this transformation no local Lagrangian will contribute to that interaction. However, the non-local Wess-Zumino term produces a term that is linear in the baryon gauge field6 :
L b -int. = b μ B μ + O ( b μ 2 ) with B μ = 1 24 π 2 ϵ μ ν ρ σ tr [ α ν α ρ α σ ] . (36) \mathcal{L}_{b\text{-int.}}=b^\mu B_\mu+O(b_\mu^2)\quad\text{with}\quad B_\mu=\frac{1}{24\pi^2}\epsilon_{\mu\nu\rho\sigma}\operatorname{tr}\left[\alpha^\nu\alpha^\rho\alpha^\sigma\right]. \tag*{(36)} L b -int. = b μ B μ + O ( b μ 2 ) with B μ = 24 π 2 1 ϵ μν ρ σ tr [ α ν α ρ α σ ] . (36)
In electrodynamics the electromagnetic field is contracted with the electromagnetic current. By analogy, B μ B_\mu B μ is the baryon number current. In particular B = ∫ d 3 r B 0 B=\int d^3r\,B_0 B = ∫ d 3 r B 0 is the baryon number. For the hedgehog configuration, Eq. (29) we get
B [ F ] = − ∫ d 3 r F ′ sin 2 F 2 π 2 = 1 π [ F ( 0 ) − F ( ∞ ) ] . (37) B[F]=-\int d^3r\,F'\frac{\sin^2F}{2\pi^2}=\frac{1}{\pi}\left[F(0)-F(\infty)\right]. \tag*{(37)} B [ F ] = − ∫ d 3 r F ′ 2 π 2 sin 2 F = π 1 [ F ( 0 ) − F ( ∞ ) ] . (37)
This finally determines F ( 0 ) = π F(0)=\pi F ( 0 ) = π for a unit baryon number configuration. The solutions for both a non-zero pion mass (μ π = 0.35 \mu_\pi=0.35 μ π = 0.35 ) and the chiral limit (μ π = 0 \mu_\pi=0 μ π = 0 ) are displayed in Figure 3 .
Figure 3. Chiral angle solving the differential equation (31).
In the chiral limit a few general statements can be made. With μ π = 0 \mu_\pi=0 μ π = 0 the integrand in Eq. (30) is free of any parameters and the classical energy of the soliton becomes E c l ≈ 72.9 f π / e E_{\rm cl}\approx72.9f_\pi/e E cl ≈ 72.9 f π / e . Furthermore, from
tr ( f π 2 α i + 1 4 e ϵ i j k α j α k ) 2 = tr ( f π 2 4 α i α i + 1 32 e 2 [ α i , α j ] 2 ± π 2 e f π B 0 ) (38) \operatorname{tr}\left(\frac{f_\pi}{2}\alpha_i+\frac{1}{4e}\epsilon_{ijk}\alpha_j\alpha_k\right)^2=\operatorname{tr}\left(\frac{f_\pi^2}{4}\alpha_i\alpha_i+\frac{1}{32e^2}[\alpha_i,\alpha_j]^2\pm\frac{\pi^2}{e}f_\pi B_0\right) \tag*{(38)} tr ( 2 f π α i + 4 e 1 ϵ ij k α j α k ) 2 = tr ( 4 f π 2 α i α i + 32 e 2 1 [ α i , α j ] 2 ± e π 2 f π B 0 ) (38)
we obtain the bound E c l ≥ 6 π 2 e f π ∣ B 0 ∣ E_{\rm cl}\ge6\frac{\pi^2}{e}f_\pi|B_0| E cl ≥ 6 e π 2 f π ∣ B 0 ∣ . This relation is often called the Bogomol’ny [30 ] bound because of its similarity to the energy bound for the ’t Hooft-Polyakov monopole [31 , 32 ] . Yet, this bound was already known to Skyrme [24 ] . The unit baryon number hedgehog solution, Eq. (29) exceeds that bound by about 20%.
Quantization
Except for the baryon number itself, the hedgehog soliton does not have appropriate quantum numbers, like spin or isospin. These will be generated in the context of canonical quantization by elevating the Noether charges for (iso)rotations to generators. To find these Noether charges we require time dependent solutions to the non-linear field equation (33) which, unfortunately, are not known. To approximate them we observe that global (iso)rotations of the chiral field are zero modes that do not cost energy. It is thus suggestive that taking those rotations to be time dependent will be a suitable approximation. For this (adiabatic) approximation we introduce collective S U ( 2 ) SU(2) S U ( 2 ) rotations as
U ( x , t ) = A ( t ) U H ( x ) A † ( t ) with A ( t ) ∈ S U ( 2 ) , (39) U(x,t)=A(t)U_H(x)A^\dagger(t)\qquad\text{with}\qquad A(t)\in SU(2), \tag*{(39)} U ( x , t ) = A ( t ) U H ( x ) A † ( t ) with A ( t ) ∈ S U ( 2 ) , (39)
where U H ( x ) U_H(x) U H ( x ) is the hedgehog configuration from Eq. (29) . At first sight, this seems as we had collective coordinates only for the iso-rotations. However, we observe
A ( t ) x ^ ⋅ τ A † ( t ) = x ^ a D a b ( A ) τ b with the S O ( 3 ) matrix D a b ( A ) = 1 2 tr [ τ a A τ b A † ] . (40) A(t)\,\hat{x}\cdot\tau\,A^\dagger(t)=\hat{x}_aD_{ab}(A)\tau_b\qquad\text{with the }SO(3)\text{ matrix}\qquad D_{ab}(A)=\frac{1}{2}\operatorname{tr}\left[\tau_aA\tau_bA^\dagger\right]. \tag*{(40)} A ( t ) x ^ ⋅ τ A † ( t ) = x ^ a D ab ( A ) τ b with the S O ( 3 ) matrix D ab ( A ) = 2 1 tr [ τ a A τ b A † ] . (40)
Hence, the collective coordinates for iso-rotations are simultaneously collective coordinates for spatial rotations. The equivalence of spatial and iso-rotations is a key feature of the hedgehog configuration. The Noether charges are associated with the (infinitesimal) transformations
U → U − ϵ ⋅ [ U , τ 2 ] + ⋯ and U → U − ϵ ′ ⋅ [ i x × ∂ , U ] + ⋯ . (41) U\to U-\epsilon\cdot\left[U,\frac{\tau}{2}\right]+\cdots\qquad\text{and}\qquad U\to U-\epsilon'\cdot\left[i x\times\partial,U\right]+\cdots. \tag*{(41)} U → U − ϵ ⋅ [ U , 2 τ ] + ⋯ and U → U − ϵ ′ ⋅ [ i x × ∂ , U ] + ⋯ . (41)
for iso- and spatial rotations, respectively. At this point it is advantageous to write the time derivative (denoted by a dot) of the collective coordinates in terms of angular velocities
Ω = − i tr ( A † ( t ) A ˙ ( t ) τ ) (42) \Omega=-i\operatorname{tr}\left(A^\dagger(t)\dot{A}(t)\tau\right) \tag*{(42)} Ω = − i tr ( A † ( t ) A ˙ ( t ) τ ) (42)
so that
U ˙ = A [ i 2 τ ⋅ Ω ′ , U H ] A † = [ i 2 τ ⋅ Ω ′ , U ] with Ω a = Ω b ′ D b a . (43) \dot{U}=A\left[\frac{i}{2}\tau\cdot\Omega',U_H\right]A^\dagger=\left[\frac{i}{2}\tau\cdot\Omega',U\right]\qquad\text{with}\qquad\Omega_a=\Omega'_bD_{ba}. \tag*{(43)} U ˙ = A [ 2 i τ ⋅ Ω ′ , U H ] A † = [ 2 i τ ⋅ Ω ′ , U ] with Ω a = Ω b ′ D ba . (43)
We immediately recognize the commutator from the iso-rotation which gives easy access to the iso-spin operator (generator):
I = − ∫ d 3 x tr { ∂ L ( U ) ∂ U ˙ ∂ U ∂ Ω ′ + h.c. } = − ∂ L [ U ] ∂ Ω ′ = − D ⋅ ∂ L [ U ] ∂ Ω . (44) I=-\int d^3x\,\operatorname{tr}\left\{\frac{\partial\mathcal{L}(U)}{\partial\dot{U}}\frac{\partial U}{\partial\Omega'}+\text{h.c.}\right\}=-\frac{\partial L[U]}{\partial\Omega'}=-D\cdot\frac{\partial L[U]}{\partial\Omega}. \tag*{(44)} I = − ∫ d 3 x tr { ∂ U ˙ ∂ L ( U ) ∂ Ω ′ ∂ U + h.c. } = − ∂ Ω ′ ∂ L [ U ] = − D ⋅ ∂ Ω ∂ L [ U ] . (44)
Similarly the hedgehog identity, Eq. (40) yields
[ i x × ∂ , U ] = − [ U , A τ 2 A † ] = − D † ⋅ [ U , τ 2 ] (45) [ix\times\partial,U]=-\left[U,A\frac{\tau}{2}A^\dagger\right]=-D^\dagger\cdot\left[U,\frac{\tau}{2}\right] \tag*{(45)} [ i x × ∂ , U ] = − [ U , A 2 τ A † ] = − D † ⋅ [ U , 2 τ ] (45)
and therefore the spin operator (generator) becomes
J = − D † ⋅ I = ∂ L [ U ] ∂ Ω . (46) J=-D^\dagger\cdot I=\frac{\partial L[U]}{\partial\Omega}. \tag*{(46)} J = − D † ⋅ I = ∂ Ω ∂ L [ U ] . (46)
We immediately observe that I 2 = J 2 I^2=J^2 I 2 = J 2 implying that the hedgehog soliton describes baryons with equal spin and iso-spin; like the nucleon (I = J = 1 2 I=J=\frac{1}{2} I = J = 2 1 ) or the Δ \Delta Δ -resonance (I = J = 3 2 I=J=\frac{3}{2} I = J = 2 3 ). At this point it is not yet clear why spin and iso-spin should be restricted to half-integer quantum numbers. Obtaining that result needs a non-trivial contribution from the Wess-Zumino term, Eq. (35) which only occurs for three flavors, cf. Section Extension to SU(3).
Global iso-rotations are parameterized by a left-multiplication of A A A with a constant S U ( 2 ) SU(2) S U ( 2 ) matrix. The hedgehog structure not only leads to Eq. (45) but also has global spatial rotations parameterized by a right-multiplication of A A A with a constant S U ( 2 ) SU(2) S U ( 2 ) matrix. Elevating I I I and J J J to the corresponding generators therefore implies (The relative sign originates from Eq. (45) .)
[ I , A ] = − τ 2 A and [ J , A ] = A τ 2 . (47) [I,A]=-\frac{\tau}{2}A \qquad\text{and} \qquad[J,A]=A\frac{\tau}{2}. \tag*{(47)} [ I , A ] = − 2 τ A and [ J , A ] = A 2 τ . (47)
Introducing Euler angles7 Φ \Phi Φ , Θ \Theta Θ and Ψ \Psi Ψ via
A ( t ) = exp ( − i Φ ( t ) τ 3 2 ) exp ( − i Θ ( t ) τ 2 2 ) exp ( − i Ψ ( t ) τ 3 2 ) (48) A(t)=\exp\left(-i\Phi(t)\frac{\tau^3}{2}\right)\exp\left(-i\Theta(t)\frac{\tau^2}{2}\right)\exp\left(-i\Psi(t)\frac{\tau^3}{2}\right) \tag*{(48)} A ( t ) = exp ( − i Φ ( t ) 2 τ 3 ) exp ( − i Θ ( t ) 2 τ 2 ) exp ( − i Ψ ( t ) 2 τ 3 ) (48)
allows one to express the generators in terms of differential operators with respect to these Euler angles. These are given in the literature, cf. the textbook [34 ] . Here it suffices to list
I z = − i ∂ ∂ Φ and J z = i ∂ ∂ Ψ . (49) I_z=-i\frac{\partial}{\partial\Phi} \qquad\text{and} \qquad J_z=i\frac{\partial}{\partial\Psi}. \tag*{(49)} I z = − i ∂ Φ ∂ and J z = i ∂ Ψ ∂ . (49)
Either set of generators is subject to an S U ( 2 ) SU(2) S U ( 2 ) algebra [ I a , I b ] = i ϵ a b c I c [I_a,I_b]=i\epsilon_{abc}I_c [ I a , I b ] = i ϵ ab c I c and [ J a , J b ] = i ϵ a b c J c [J_a,J_b]=i\epsilon_{abc}J_c [ J a , J b ] = i ϵ ab c J c .
Lastly, it remains to find the Ω \Omega Ω dependence of the Lagrange function. This is straightforward: we substitute the rotating hedgehog configuration, Eq. (39) into the Lagrangian, Eq. (28) and integrate over space:
L [ U ] = 1 2 α 2 [ F ] Ω ⋅ Ω − E c l [ F ] with α 2 [ F ] = 8 π 3 f π e 3 ∫ 0 ∞ d x x 2 sin 2 F [ 1 + ( F ′ 2 + sin 2 F x 2 ) ] . (50) L[U]=\frac{1}{2}\alpha^2[F]\,\boldsymbol{\Omega}\cdot\boldsymbol{\Omega}-E_{\mathrm{cl}}[F]
\qquad\text{with} \qquad
\alpha^2[F]=\frac{8\pi}{3f_\pi e^3}\int_0^\infty dx\,x^2\sin^2F\left[1+\left(F'^2+\frac{\sin^2F}{x^2}\right)\right]. \tag*{(50)} L [ U ] = 2 1 α 2 [ F ] Ω ⋅ Ω − E cl [ F ] with α 2 [ F ] = 3 f π e 3 8 π ∫ 0 ∞ d x x 2 sin 2 F [ 1 + ( F ′2 + x 2 sin 2 F ) ] . (50)
Again, we can easily read off that the moment of inertia scales like α 2 = O ( N C ) \alpha^2=O(N_C) α 2 = O ( N C ) and in the chiral limit (μ π = 0 \mu_\pi=0 μ π = 0 ) we have the universal result α 2 = 51.2 / ( e 3 f π ) \alpha^2=51.2/(e^3f_\pi) α 2 = 51.2/ ( e 3 f π ) . Inverting Ω = 1 α 2 J \boldsymbol{\Omega}=\frac{1}{\alpha^2}J Ω = α 2 1 J we are in the position to write down a baryon mass formula
E I = J = J ( J + 1 ) 2 α 2 + E c l ⟹ E N = 3 8 α 2 + E c l and E Δ = 15 8 α 2 + E c l , (51) E_{I=J}=\frac{J(J+1)}{2\alpha^2}+E_{\mathrm{cl}}
\qquad\Longrightarrow\qquad
E_N=\frac{3}{8\alpha^2}+E_{\mathrm{cl}}
\qquad\text{and}\qquad
E_\Delta=\frac{15}{8\alpha^2}+E_{\mathrm{cl}}, \tag*{(51)} E I = J = 2 α 2 J ( J + 1 ) + E cl ⟹ E N = 8 α 2 3 + E cl and E Δ = 8 α 2 15 + E cl , (51)
where we used that nucleon and Δ \Delta Δ -resonance have (iso)spin 1 2 \frac{1}{2} 2 1 and 3 2 \frac{3}{2} 2 3 , respectively. Obviously, the correction from generating spin and isospin quantum numbers is O ( 1 / N C ) O(1/N_C) O ( 1/ N C ) . Using the above universal results for E c l E_{\mathrm{cl}} E cl and α 2 \alpha^2 α 2 when μ π = 0 \mu_\pi=0 μ π = 0 would then predict f π ≈ 64.5 f_\pi\approx64.5 f π ≈ 64.5 MeV and e ≈ 5.4 e\approx5.4 e ≈ 5.4 . Stated conversely, when using the empirical value for the pion decay constant from Section chiral Lagrangian , the Skyrme model overestimates the nucleon mass by 30% or more. Taking the actual pion mass does not alter that conclusion significantly [35 ] . Adjusting the pion decay constant was the original approach in Ref. [33 ] . Nowadays one considers quantum corrects to the classical mass that are O ( N C 0 ) O(N_C^0) O ( N C 0 ) , and neither depend on spin nor isospin, to give a substantial negative contribution to the energy [19 ] . It is therefore customary to keep f π = 93 f_\pi=93 f π = 93 MeV as well as m π = 135 m_\pi=135 m π = 135 MeV and tune e ≈ 4.25 e\approx4.25 e ≈ 4.25 to reproduce E Δ − E N = 293 E_\Delta-E_N=293 E Δ − E N = 293 MeV; even though these corrections are not fully under control since the model is not renormalizable. Here one may also remark that the chiral-limit decay constant is smaller than the physical value [36 ] .
Static properties
To compute static properties we need to determine the symmetry currents from vector and axial-vector transformations. That is, we seek the Skyrme model analog of Eq. (4) . A convenient method (already mentioned above) is to upgrade these global symmetries to local ones by extending the Skyrme model action with appropriate external gauge fields (e.g. the gauge fields of the electroweak interactions). The Noether currents are subsequently read off as the objects which couple linearly to these gauge fields. This procedure results in Lorentz covariant expressions for the currents and is especially appropriate for the Wess-Zumino term (34) because this non-local term can only be made gauge invariant by a trial and error type procedure [27 , 29 , 37 ] . Substituting the rotating hedgehog, Eq. (39) into these expressions yields the densities
1 3 V 0 0 = 1 2 b ( r ) , 1 3 V i 0 = 1 2 b ( r ) ϵ i j k Ω j x k , V 0 a = − 2 3 v ( r ) D a i Ω i and V i a = v ( r ) r 2 ϵ i j k x j D a k . (52) \frac{1}{3}V_0^0=\frac{1}{2}b(r), \qquad
\frac{1}{3}V_i^0=\frac{1}{2}b(r)\epsilon_{ijk}\Omega_jx_k, \qquad
V_0^a=-\frac{2}{3}v(r)D_{ai}\Omega_i
\qquad\text{and}\qquad
V_i^a=\frac{v(r)}{r^2}\epsilon_{ijk}x_jD_{ak}. \tag*{(52)} 3 1 V 0 0 = 2 1 b ( r ) , 3 1 V i 0 = 2 1 b ( r ) ϵ ij k Ω j x k , V 0 a = − 3 2 v ( r ) D ai Ω i and V i a = r 2 v ( r ) ϵ ij k x j D ak . (52)
Table 1. Predictions for nucleon static properties (magnetic moments, electric radii and axial vector charge) in the Skyrme model compared to experimental data [9 ] . Two parameter sets are considered: (A: f π = 93 f_\pi= 93 f π = 93 MeV e = 4.25 e = 4.25 e = 4.25 ) and (B: f π = 54 f_\pi= 54 f π = 54 MeV e = 4.84 e = 4.84 e = 4.84 [35 ] ). Either case has the physical pion mass m π = 135 m_\pi= 135 m π = 135 MeV.
with (here primes are derivatives with respect to r r r )
b ( r ) = − F ′ sin 2 F 2 π 2 and v ( r ) = sin 2 F [ f π 2 + 1 e 2 ( F ′ 2 + sin 2 F r 2 ) ] (53) b(r)=-F'\frac{\sin^2 F}{2\pi^2}
\qquad\text{and}\qquad
v(r)=\sin^2 F\left[f_\pi^2+\frac{1}{e^2}\left(F'^2+\frac{\sin^2 F}{r^2}\right)\right] \tag*{(53)} b ( r ) = − F ′ 2 π 2 sin 2 F and v ( r ) = sin 2 F [ f π 2 + e 2 1 ( F ′2 + r 2 sin 2 F ) ] (53)
for the vector currents. Subscripts are Lorentz labels while the superscripts refer to the flavor, i.e., isospin quantum number. Similarly the non-zero elements of the axial vector current are
A i a = [ a 1 ( r ) δ i k + a 2 ( r ) x ^ i x ^ k ] D a k (54) A_i^a=\left[a_1(r)\delta_{ik}+a_2(r)\hat{x}_i\hat{x}_k\right]D_{ak} \tag*{(54)} A i a = [ a 1 ( r ) δ ik + a 2 ( r ) x ^ i x ^ k ] D ak (54)
with the densities
a 1 ( r ) = sin 2 F 2 r [ f π 2 + 1 e 2 ( F ′ 2 + sin 2 F r 2 ) ] and a 2 ( r ) = − a 1 ( r ) + F ′ [ f π 2 + 2 sin 2 F e 2 r 2 ] (55) a_1(r)=\frac{\sin^2 F}{2r}\left[f_\pi^2+\frac{1}{e^2}\left(F'^2+\frac{\sin^2 F}{r^2}\right)\right]
\qquad\text{and}\qquad
a_2(r)=-a_1(r)+F'\left[f_\pi^2+\frac{2\sin^2 F}{e^2r^2}\right] \tag*{(55)} a 1 ( r ) = 2 r sin 2 F [ f π 2 + e 2 1 ( F ′2 + r 2 sin 2 F ) ] and a 2 ( r ) = − a 1 ( r ) + F ′ [ f π 2 + e 2 r 2 2 sin 2 F ] (55)
Here, and in Eq. (53) , primes denote derivatives with respect to r r r . While the conservation of the vector current ∂ μ V μ a = 0 \partial^\mu V_\mu^a=0 ∂ μ V μ a = 0 follows from the structure of the currents in Eq. (52) , the axial vector current analog ∂ μ A μ a = m π 2 x ^ a sin F \partial^\mu A_\mu^a=m_\pi^2\hat{x}^a\sin F ∂ μ A μ a = m π 2 x ^ a sin F results from Eq. (31) . The fact that it vanishes in the chiral limit is known as PCAC (partially conserved axial vector current).
We have already discussed how to relate some of the functions of the collective coordinates to spin and isospin in the preceding subsection. To compute matrix elements of D a i D_{ai} D ai we either use Wigner-D D D functions as wave-functions, D I 3 , − J 3 ( I = J ) ∗ ( Φ , Θ , Ψ ) D^{(I=J)*}_{I_3,-J_3}(\Phi,\Theta,\Psi) D I 3 , − J 3 ( I = J ) ∗ ( Φ , Θ , Ψ ) and integrate over the Euler angles, or note that the left index (a a a ) is isospin, the right one (i i i ) is spin, hence ⟨ D a i ⟩ ∼ I a J i \langle D_{ai}\rangle\sim I_aJ_i ⟨ D ai ⟩ ∼ I a J i . The remaining reduced matrix element is easily obtained from the expectation value ⟨ D a i J i ⟩ = − ⟨ I a ⟩ \langle D_{ai}J_i\rangle=-\langle I_a\rangle ⟨ D ai J i ⟩ = − ⟨ I a ⟩ . We summarize
Ω i = 1 α 2 J i , D a i Ω i = − 1 α 2 I a and ⟨ D a i ⟩ = − 1 J ( J + 1 ) I a J i . (56) \Omega_i=\frac{1}{\alpha^2}J_i,\qquad D_{ai}\Omega_i=-\frac{1}{\alpha^2}I_a
\qquad\text{and}\qquad
\langle D_{ai}\rangle=-\frac{1}{J(J+1)}I_aJ_i. \tag*{(56)} Ω i = α 2 1 J i , D ai Ω i = − α 2 1 I a and ⟨ D ai ⟩ = − J ( J + 1 ) 1 I a J i . (56)
With ∫ d 3 r v ( r ) = 3 2 α 2 \int d^3r\,v(r)=\frac{3}{2}\alpha^2 ∫ d 3 r v ( r ) = 2 3 α 2 , we immediately get the electromagnetic charges8 ^8 8 ⟨ Q ⟩ = ∫ d 3 r V 0 e m . = ∫ d 3 r ( 1 3 V 0 0 + V 0 3 ) = 1 2 + I 3 \langle Q\rangle=\int d^3r\,V_0^{\mathrm{em.}}=\int d^3r\left(\frac{1}{3}V_0^0+V_0^3\right)=\frac{1}{2}+I_3 ⟨ Q ⟩ = ∫ d 3 r V 0 em. = ∫ d 3 r ( 3 1 V 0 0 + V 0 3 ) = 2 1 + I 3 . We are now prepared to compute nucleon static properties. As an explicit example we present the magnetic moment
μ N μ n . m . = 1 μ n . m . ⟨ N J 3 = 1 2 ∣ ∫ d 3 x ( x × V e m . ) 3 ∣ N J 3 = 1 2 ⟩ = 2 π 3 M N ∫ 0 ∞ d r r 2 [ 2 3 v ( r ) + r 2 2 α 2 b ( r ) ] (57) \frac{\mu_N}{\mu_{\mathrm{n.m.}}}=\frac{1}{\mu_{\mathrm{n.m.}}}\left\langle NJ_3=\frac{1}{2}\left|\int d^3x\left(\boldsymbol{x}\times\boldsymbol{V}^{\mathrm{em.}}\right)_3\right|NJ_3=\frac{1}{2}\right\rangle=\frac{2\pi}{3}M_N\int_0^\infty dr\,r^2\left[\frac{2}{3}v(r)+\frac{r^2}{2\alpha^2}b(r)\right] \tag*{(57)} μ n.m. μ N = μ n.m. 1 ⟨ N J 3 = 2 1 ∫ d 3 x ( x × V em. ) 3 N J 3 = 2 1 ⟩ = 3 2 π M N ∫ 0 ∞ d r r 2 [ 3 2 v ( r ) + 2 α 2 r 2 b ( r ) ] (57)
typically measured in nucleon magnetons μ n . m . = e ℏ 2 M N \mu_{\mathrm{n.m.}}=\frac{e\hbar}{2M_N} μ n.m. = 2 M N e ℏ . Here M N = 939 M_N=939 M N = 939 MeV is the physical nucleon mass not the model prediction because it enters solely via μ n . m . \mu_{\mathrm{n.m.}} μ n.m. . The two signs in Eq. (57) refer to proton (+) and neutron (−). Numerical results for two sets of model parameters are compared to data in table 1, which also contains the axial vector charge g a = 2 ⟨ p ∣ A 3 3 ∣ p ⟩ g_a=2\langle p|A_3^3|p\rangle g a = 2 ⟨ p ∣ A 3 3 ∣ p ⟩ . Both sets reproduce the empirical Δ \Delta Δ -nucleon mass difference. While set A uses the physical pion decay constant, set B tunes it to match the nucleon mass. We see that, in either case, the model predictions are too low in magnitude by up to 30%. The main cause is the too narrow chiral angle, needed to get α 2 \alpha^2 α 2 small enough at the expense of small radii.
Magnetic radii (not listed in table 1) have an additional factor r 2 r^2 r 2 under the integral, Eq. (57) , and in view of Eq. (32) require m π ≠ 0 m_\pi\ne0 m π = 0 to be finite.
An important result is A 0 0 = 0 A_0^0=0 A 0 0 = 0 , i.e., the axial singlet current vanishes identically in the Skyrme model. This has been viewed as an elegant explanation of the proton spin puzzle [38 ] and ignited a kind of revival (after most of the static properties had been calculated) of chiral soliton models. Among other topics we will get back to this one in the Section Vector mesons .
The computation of momentum dependent form factors is also possible. It requires an additional time-dependent collective coordinate for the position of the soliton, R ( t ) R(t) R ( t ) :
U ( x , t ) = A ( t ) U H ( x − R ( t ) ) A † ( t ) . (58) U(x,t)=A(t)U_H(x-R(t))A^\dagger(t). \tag*{(58)} U ( x , t ) = A ( t ) U H ( x − R ( t )) A † ( t ) . (58)
Its canonical quantization leads to a plane wave wave-function with momentum P P P . The form factors are then obtained as Fourier transforms of the above listed densities. In fact, the above calculation just corresponds to the P = 0 P=0 P = 0 case. In Ref. [39 ] the calculational techniques are detailed and the Skyrme model predictions are compared to the standard dipole parameterization of experimental data based on a vector meson dominance picture.
As an alternative to stabilize the soliton, a sixth order term has been considered [40 ] : L 6 = − e 6 2 4 B μ B μ \mathcal{L}_6=-\frac{e_6^2}{4}B_\mu B^\mu L 6 = − 4 e 6 2 B μ B μ . No substantial differences have been encountered.
Pion-nucleon scattering
As for the simple kink model in Section Solitons in field theory , meson-baryon scattering is analyzed in terms of small amplitude fluctuations about the soliton. A suggestive parameterization is [41 , 42 ]
U ( x , t ) = A exp ( i τ ⋅ x ^ F ( r ) + i τ ⋅ η ( x , t ) ) A † . (59) U(x,t)=A\exp\left(i\boldsymbol{\tau}\cdot\hat{x}F(r)+i\boldsymbol{\tau}\cdot\boldsymbol{\eta}(x,t)\right)A^\dagger. \tag*{(59)} U ( x , t ) = A exp ( i τ ⋅ x ^ F ( r ) + i τ ⋅ η ( x , t ) ) A † . (59)
In the adiabatic approximation the collective rotation A A A is treated as time-independent and the linearized field equations
M i j μ ν ( x ) ∂ μ ∂ ν η j ( x , t ) + V i j ( x ) η j ( x , t ) = 0 (60) M_{ij}^{\mu\nu}(x)\partial_\mu\partial_\nu\eta_j(x,t)+V_{ij}(x)\eta_j(x,t)=0 \tag*{(60)} M ij μν ( x ) ∂ μ ∂ ν η j ( x , t ) + V ij ( x ) η j ( x , t ) = 0 (60)
reduce to Klein-Gordon equations asymptotically, i.e., r → ∞ r\to\infty r → ∞ : M i j μ ν ( x ) → M i j ( 0 ) μ ν = g μ ν δ i j M_{ij}^{\mu\nu}(x)\to M_{ij}^{(0)\mu\nu}=g^{\mu\nu}\delta_{ij} M ij μν ( x ) → M ij ( 0 ) μν = g μν δ ij and V i j ( x ) → V i j ( 0 ) = m π 2 δ i j V_{ij}(x)\to V_{ij}^{(0)}=m_\pi^2\delta_{ij} V ij ( x ) → V ij ( 0 ) = m π 2 δ ij . The space dependence of the matrices M i j μ ν ( x ) M_{ij}^{\mu\nu}(x) M ij μν ( x ) and V i j ( x ) V_{ij}(x) V ij ( x ) is generated by the soliton, U H ( x ) U_H(x) U H ( x ) . The fluctuations have the partial wave decomposition
η ( x , t ) = e − i ω t ∑ G L M η G L ( r , ω ) Y G L M ( x ^ ) , (61) \eta(x,t)=e^{-i\omega t}\sum_{GLM}\eta_{GL}(r,\omega)Y_{GLM}(\hat{x}), \tag*{(61)} η ( x , t ) = e − iω t G L M ∑ η G L ( r , ω ) Y G L M ( x ^ ) , (61)
where Y G L M Y_{GLM} Y G L M are vector spherical harmonics for the conserved grand spin, G = L + I G=L+I G = L + I which is the operator sum of the pion’s orbital angular momentum and its isospin. The conservation of grand spin results from the hedgehog being invariant under a combination of spatial and iso-rotations, cf. Eq. (45) . The time dependence factorizes because the soliton is static, at least in the adiabatic approximation.
The scattering amplitudes are extracted from the asymptotic behavior of the regular radial functions in Eq. (61) . The resulting, so-called intrinsic scattering matrix elements are labeled as S ~ ℓ , ℓ ′ G \widetilde{S}^{G}_{\ell,\ell'} S ℓ , ℓ ′ G , where ℓ , ℓ ′ ∈ [ G − 1 , G , G + 1 ] \ell,\ell'\in[G-1,G,G+1] ℓ , ℓ ′ ∈ [ G − 1 , G , G + 1 ] are the possible pion angular momenta for prescribed grand spin G G G . The physical S S S -matrix elements are finally obtained from the recoupling scheme [41 ] .9
S ℓ ( 2 I ) ( 2 J ) = ∑ G ζ ζ ′ S ~ ℓ , ℓ ′ G , with the recoupling coefficient ζ = ( − 1 ) ℓ + 1 2 + J + G [ 2 ( 2 G + 1 ) ] { I 1 1 2 ℓ J G } . (62) S_{\ell(2I)(2J)}=\sum_G\zeta\zeta'\widetilde{S}^{G}_{\ell,\ell'},\qquad\text{with the recoupling coefficient}\qquad\zeta=(-1)^{\ell+\frac{1}{2}+J+G}\sqrt{[2(2G+1)]}\left\{\begin{matrix} I & 1 & \frac{1}{2} \\ \ell& J & G \end{matrix}\right\}. \tag*{(62)} S ℓ ( 2 I ) ( 2 J ) = G ∑ ζ ζ ′ S ℓ , ℓ ′ G , with the recoupling coefficient ζ = ( − 1 ) ℓ + 2 1 + J + G [ 2 ( 2 G + 1 )] { I ℓ 1 J 2 1 G } . (62)
Here, J J J and I I I are total spin and isospin. The entry ’1’ arises from the pion isospin while ’1 2 \frac{1}{2} 2 1 ’ reflects the nucleon spin. The interesting observation is that there are more elements of the physical S S S -matrix than there are for the intrinsic one, S ~ \widetilde{S} S . The latter can thus be eliminated in favor of linear relations among the physical S S S -matrix elements. For the elastic pion-nucleon scattering this predicts linear dependences of the I = 3 2 I=\frac{3}{2} I = 2 3 and I = 1 2 I=\frac{1}{2} I = 2 1 scattering amplitudes [41 ] ,
2 ( 2 ℓ + 1 ) S ℓ ( 3 ) ( 2 ℓ + 1 ) = 3 ℓ S ℓ ( 1 ) ( 2 ℓ + 1 ) + ( ℓ + 2 ) S ℓ ( 1 ) ( 2 ℓ + 1 ) (63) 2(2\ell+1)S_{\ell(3)(2\ell+1)}=3\ell S_{\ell(1)(2\ell+1)}+(\ell+2)S_{\ell(1)(2\ell+1)} \tag*{(63)} 2 ( 2 ℓ + 1 ) S ℓ ( 3 ) ( 2 ℓ + 1 ) = 3 ℓ S ℓ ( 1 ) ( 2 ℓ + 1 ) + ( ℓ + 2 ) S ℓ ( 1 ) ( 2 ℓ + 1 ) (63)
and
2 ( 2 ℓ + 1 ) S ℓ ( 3 ) ( 2 ℓ − 1 ) = ( ℓ − 1 ) S ℓ ( 1 ) ( 2 ℓ − 1 ) + 3 ( ℓ + 2 ) S ℓ ( 1 ) ( 2 ℓ − 1 ) . (64) 2(2\ell+1)S_{\ell(3)(2\ell-1)}=(\ell-1)S_{\ell(1)(2\ell-1)}+3(\ell+2)S_{\ell(1)(2\ell-1)}. \tag*{(64)} 2 ( 2 ℓ + 1 ) S ℓ ( 3 ) ( 2 ℓ − 1 ) = ( ℓ − 1 ) S ℓ ( 1 ) ( 2 ℓ − 1 ) + 3 ( ℓ + 2 ) S ℓ ( 1 ) ( 2 ℓ − 1 ) . (64)
These relations are model independent in the sense that they are valid irrespective of the details of the Lagrangian, as long as scattering is explored in the adiabatic approximation. They are well satisfied by data for ℓ ≥ 3 \ell\geq3 ℓ ≥ 3 but for lower angular momenta the approximation is not adequate. This short-coming is closely related to the Yukawa problem of soliton models: by definition, there is no term linear in the fluctuations about the soliton and thus there is no Yukawa coupling to a resonance. A linear term nevertheless emerges in the context of the collective coordinate quantization because Eq. (39) is not an exact solution. Numerous attempts have been made to solve this problem, cf. Ref. [44 ] and references therein. Another problem arises from the fact that in the Skyrme model the metric tensor, Eq. (60) has M i j μ ν ≠ M i j μ ν M_{ij}^{\mu\nu}\neq M_{ij}^{\mu\nu} M ij μν = M ij μν . As a result, the phase shifts do not saturate at large energies and the identification of genuine resonances is indistinct. That problem, however, has a solution and we will reflect on it in Section Vector mesons .
Another important result arises from Eq. (60) : for the parameterization, Eq. (61) all energy eigenvalues ω \omega ω are found to be real. In fact the lowest eigenvalues are the zero modes associated with translations and (iso)rotations [42 ] . Hence the solution arising from the ansatz , Eq. (29) is stable, at least locally.
Extension to S U ( 3 ) SU(3) S U ( 3 )
Empirically chiral symmetry can be considered to be approximately realized when pseudoscalar and vector mesons with the same quark content differ significantly in mass. For the up and down flavors this is obviously the case: m π = 135 M e V ≪ 770 M e V ≃ m ρ m_\pi=135\,\mathrm{MeV}\ll770\,\mathrm{MeV}\simeq m_\rho m π = 135 MeV ≪ 770 MeV ≃ m ρ . For mesons containing an up or a down and a strange (anti)quark, it is approximately true: m K = 495 M e V < 890 M e V ≃ m K ∗ m_K=495\,\mathrm{MeV}<890\,\mathrm{MeV}\simeq m_{K^*} m K = 495 MeV < 890 MeV ≃ m K ∗ [9 ] . Therefore it is legitimate to extend the Skyrme model to favor S U ( 3 ) SU(3) S U ( 3 ) . Without flavor symmetry breaking, any embedding of the hedgehog, Eq. (29) would be a solution to the field equations. Including (some) explicit breaking as measured by m K > m π m_K>m_\pi m K > m π has a larger classical energy when the hedgehog has non-zero kaon components. Whence the initial configuration in the three flavor model is [28 , 45 , 46 ]
U H ( x ) = ( exp ( i τ ⋅ x ^ F ( r ) ) 0 0 1 ) , (65) U_H(x)=\begin{pmatrix}
\exp\left(i\boldsymbol{\tau}\cdot\hat{x}F(r)\right) & 0\\
0 & 1
\end{pmatrix}, \tag*{(65)} U H ( x ) = ( exp ( i τ ⋅ x ^ F ( r ) ) 0 0 1 ) , (65)
where, for the Skyrme model case, F ( r ) F(r) F ( r ) is again the solution to the equation of motion, Eq. (31) which is displayed in Figure 3 . We introduce collective coordinates for all zero modes in the S U ( 3 ) SU(3) S U ( 3 ) symmetric model
U ( x , t ) = A ( t ) U H ( x ) A † ( t ) (66) U(x,t)=A(t)U_{\mathrm{H}}(x)A^\dagger(t) \tag*{(66)} U ( x , t ) = A ( t ) U H ( x ) A † ( t ) (66)
to parameterize time dependent configurations as in Eq. (66) . Here the soliton rotates without deformation (i.e., rigidly) in flavor space. Therefore treatments based on Eq. (66) are commonly summarized as the rigid rotator approach (RRA). When we substitute this ansatz into the sum ∫ d 4 x L S k + Γ W Z \int d^4x\,\mathcal{L}_{\mathrm{Sk}}+\Gamma_{\mathrm{WZ}} ∫ d 4 x L Sk + Γ WZ , we obtain the Lagrange function for the collective coordinates
L ( A , Ω a ) = − E c l + 1 2 α 2 ∑ i = 1 3 Ω i 2 + 1 2 β 2 ∑ α = 4 7 Ω α 2 − N C B 2 3 Ω 8 (67) L(A,\Omega_a)=-E_{\mathrm{cl}}+\frac{1}{2}\alpha^2\sum_{i=1}^{3}\Omega_i^2+\frac{1}{2}\beta^2\sum_{\alpha=4}^{7}\Omega_\alpha^2-\frac{N_{\mathrm{C}}B}{2\sqrt{3}}\Omega_8 \tag*{(67)} L ( A , Ω a ) = − E cl + 2 1 α 2 i = 1 ∑ 3 Ω i 2 + 2 1 β 2 α = 4 ∑ 7 Ω α 2 − 2 3 N C B Ω 8 (67)
with the three flavor generalization of Eq. (42) : Ω a = − i tr [ A † ( t ) A ˙ ( t ) λ a ] \Omega_a=-i\operatorname{tr}\left[A^\dagger(t)\dot{A}(t)\lambda_a\right] Ω a = − i tr [ A † ( t ) A ˙ ( t ) λ a ] . A second moment of inertia
β 2 = π e 3 f π ∫ 0 ∞ d x x 2 sin 2 F 2 [ 4 + ( F ′ 2 + 2 x 2 sin 2 F ) ] (68) \beta^2=\frac{\pi}{e^3f_\pi}\int_0^\infty dx\,x^2\sin^2\frac{F}{2}\left[4+\left(F^{\prime2}+\frac{2}{x^2}\sin^2F\right)\right] \tag*{(68)} β 2 = e 3 f π π ∫ 0 ∞ d x x 2 sin 2 2 F [ 4 + ( F ′2 + x 2 2 sin 2 F ) ] (68)
emerges because the embedding, Eq. (65) breaks the symmetry for rotations into strangeness direction. It is O ( N C ) O(N_{\mathrm{C}}) O ( N C ) and typical model results for this inertia parameter are about 4 G e V − 1 4\,\mathrm{GeV}^{-1} 4 GeV − 1 . In the chiral limit we have the universal result β 2 = 56.5 / ( e 3 f π ) \beta^2=56.5/(e^3f_\pi) β 2 = 56.5/ ( e 3 f π ) . Even though [ U H , λ 8 ] = 0 [U_{\mathrm{H}},\lambda_8]=0 [ U H , λ 8 ] = 0 , Ω 8 \Omega_8 Ω 8 appears in Eq. (67) because of the non-local nature of the Wess-Zumino term. Quantization of this system generalizes Eq. (46) by specifying (intrinsic) S U ( 3 ) SU(3) S U ( 3 ) generators
R a = − ∂ L ∂ Ω a = { − α 2 Ω a = − J a , a = 1 , 2 , 3 − β 2 Ω a , a = 4 , … , 7 N C B 2 3 , a = 8. (69) R_a=-\frac{\partial L}{\partial\Omega_a}=
\begin{cases}
-\alpha^2\Omega_a=-J_a, & a=1,2,3\\
-\beta^2\Omega_a, & a=4,\ldots,7\\
\frac{N_{\mathrm{C}}B}{2\sqrt{3}}, & a=8.
\end{cases} \tag*{(69)} R a = − ∂ Ω a ∂ L = ⎩ ⎨ ⎧ − α 2 Ω a = − J a , − β 2 Ω a , 2 3 N C B , a = 1 , 2 , 3 a = 4 , … , 7 a = 8. (69)
Quantization imposes the commutation relations [ R a , R b ] = − i f a b c R c [R_a,R_b]=-if_{abc}R_c [ R a , R b ] = − i f ab c R c , with S U ( 3 ) SU(3) S U ( 3 ) structure constants f a b c f_{abc} f ab c . The fact that the first three generators are identified with the spin operator is a mere consequence of the hedgehog structure that identifies spin and isospin rotations. We easily find the Hamiltonian from the Legendre transformation
H = E c l + 1 2 α 2 J 2 + 1 2 β 2 ∑ α = 4 7 R α 2 = E c l + 1 2 ( 1 α 2 − 1 β 2 ) J 2 + 1 2 β 2 C 2 [ S U ( 3 ) ] − N C 2 B 2 24 β 2 (70) H=E_{\mathrm{cl}}+\frac{1}{2\alpha^2}\boldsymbol{J}^2+\frac{1}{2\beta^2}\sum_{\alpha=4}^{7}R_\alpha^2=E_{\mathrm{cl}}+\frac{1}{2}\left(\frac{1}{\alpha^2}-\frac{1}{\beta^2}\right)\boldsymbol{J}^2+\frac{1}{2\beta^2}C_2[SU(3)]-\frac{N_{\mathrm{C}}^2B^2}{24\beta^2} \tag*{(70)} H = E cl + 2 α 2 1 J 2 + 2 β 2 1 α = 4 ∑ 7 R α 2 = E cl + 2 1 ( α 2 1 − β 2 1 ) J 2 + 2 β 2 1 C 2 [ S U ( 3 )] − 24 β 2 N C 2 B 2 (70)
where C 2 [ S U ( 3 ) ] = ∑ a = 1 8 R a 2 C_2[SU(3)]=\sum_{a=1}^{8}R_a^2 C 2 [ S U ( 3 )] = ∑ a = 1 8 R a 2 is the quadratic Casimir operator of S U ( 3 ) SU(3) S U ( 3 ) . The last term in Eq. (70) does not affect mass differences as long as only unit baryon number states are considered. However, we need to find the eigenstates subject to the constraint Y R = 2 3 R 8 = N C B 3 Y_R=\frac{2}{\sqrt{3}}R_8=\frac{N_{\mathrm{C}}B}{3} Y R = 3 2 R 8 = 3 N C B . This can be done by a detailed analysis of S U ( 3 ) SU(3) S U ( 3 ) representations. A more intuitive derivation notes that the well-known Gell-Mann-Nishijima relation for the electric charges has the analog Q R = R 3 + 1 2 Y R = − J 3 + 1 2 Y R Q_R=R_3+\frac{1}{2}Y_R=-J_3+\frac{1}{2}Y_R Q R = R 3 + 2 1 Y R = − J 3 + 2 1 Y R . The quark model tells us that possible eigenvalues are Q R = 0 , ± 1 3 , ± 2 3 , ± 1 , 1 3 , … Q_R=0,\pm\frac{1}{3},\pm\frac{2}{3},\pm1,\frac{1}{3},\ldots Q R = 0 , ± 3 1 , ± 3 2 , ± 1 , 3 1 , … , implying that half-integer charges are not allowed. For Y R = 1 Y_R=1 Y R = 1 the constraint can thus only be consistently accommodated with J 3 J_3 J 3 being half-integer. Hence for N C = 3 N_{\mathrm{C}}=3 N C = 3 baryons indeed have half-integer spin [28 ] . Furthermore, for Y R = 1 Y_R=1 Y R = 1 the allowed low dimensional S U ( 3 ) SU(3) S U ( 3 ) representations are the octet and the decuplet, containing the established spin 1 2 \frac{1}{2} 2 1 and 3 2 \frac{3}{2} 2 3 baryons, respectively. The corresponding C 2 [ S U ( 2 ) ] C_2[SU(2)] C 2 [ S U ( 2 )] eigenvalues are three and six, so that the mass difference between the spin 1 2 \frac{1}{2} 2 1 and 3 2 \frac{3}{2} 2 3 baryons is M 10 − M 8 = 3 2 α 2 M_{10}-M_8=\frac{3}{2\alpha^2} M 10 − M 8 = 2 α 2 3 , as in the two flavor version.
Flavor symmetry breaking terms that transform like ( L R † + h . c . ) (LR^\dagger+\mathrm{h.c.}) ( L R † + h.c. ) are added to the chiral Lagrangian to account for different masses and decay constants f K ≈ 1.22 f π f_K\approx1.22f_\pi f K ≈ 1.22 f π :
L s b = f π 2 m π 2 + 2 f K 2 m K 2 12 tr [ U + U † − 2 ] + 3 f π 2 m π 2 − f K 2 m K 2 6 tr [ λ 8 ( U + U † ) ] + f K 2 − f π 2 12 tr [ ( 1 − 3 λ 8 ) ( ∂ μ U ∂ μ U † + U † ∂ μ U ∂ μ U † ) ] (71) \mathcal{L}_{\mathrm{sb}}=\frac{f_\pi^2m_\pi^2+2f_K^2m_K^2}{12}\operatorname{tr}\left[U+U^\dagger-2\right]+\sqrt{3}\frac{f_\pi^2m_\pi^2-f_K^2m_K^2}{6}\operatorname{tr}\left[\lambda_8\left(U+U^\dagger\right)\right]+\frac{f_K^2-f_\pi^2}{12}\operatorname{tr}\left[(1-\sqrt{3}\lambda_8)\left(\partial_\mu U\partial^\mu U^\dagger+U^\dagger\partial_\mu U\partial^\mu U^\dagger\right)\right] \tag*{(71)} L sb = 12 f π 2 m π 2 + 2 f K 2 m K 2 tr [ U + U † − 2 ] + 3 6 f π 2 m π 2 − f K 2 m K 2 tr [ λ 8 ( U + U † ) ] + 12 f K 2 − f π 2 tr [ ( 1 − 3 λ 8 ) ( ∂ μ U ∂ μ U † + U † ∂ μ U ∂ μ U † ) ] (71)
Generalizing the definition in Eq. (40) to D a b ( A ) = 1 2 tr [ λ a A λ b A † ] D_{ab}(A)=\frac{1}{2}\operatorname{tr}\left[\lambda_aA\lambda_bA^\dagger\right] D ab ( A ) = 2 1 tr [ λ a A λ b A † ] , L s b \mathcal{L}_{\mathrm{sb}} L sb adds
H s b = 1 2 γ ( 1 − D 88 ( A ) ) (72) H_{\mathrm{sb}}=\frac{1}{2}\gamma(1-D_{88}(A)) \tag*{(72)} H sb = 2 1 γ ( 1 − D 88 ( A )) (72)
with
γ = 4 3 ∫ d 3 r { ( m K 2 f K 2 − m π 2 f π 2 ) ( 1 − cos F ) + 1 2 ( f K 2 − f π 2 ) cos F ( F ′ 2 + 2 sin 2 F r 2 ) } . \gamma=\frac{4}{3}\int d^3r\left\{(m_K^2f_K^2-m_\pi^2f_\pi^2)(1-\cos F)+\frac{1}{2}(f_K^2-f_\pi^2)\cos F\left(F^{\prime2}+\frac{2\sin^2F}{r^2}\right)\right\}. γ = 3 4 ∫ d 3 r { ( m K 2 f K 2 − m π 2 f π 2 ) ( 1 − cos F ) + 2 1 ( f K 2 − f π 2 ) cos F ( F ′2 + r 2 2 sin 2 F ) } .
to the Hamilton operator, Eq. (70) . The constraint for R 8 R_8 R 8 is not effected. Using the S U ( 3 ) SU(3) S U ( 3 ) Clebsch-Gordan coefficients [47 ] H s b H_{\mathrm{sb}} H sb can be treated in stationary perturbation theory [45 , 48 ] . At leading order this yields the model independent predictions
2 ( M N + M Ξ ) = 3 M Λ + M Σ , M Δ − M Σ ∗ = M Σ ∗ − M Ξ ∗ = M Ξ ∗ − M Ω and M Ξ ∗ − M Σ ∗ + M N = 1 8 ( 11 M Λ − 3 M Σ ) . (73) 2(M_N+M_\Xi)=3M_\Lambda+M_\Sigma,\qquad M_\Delta-M_{\Sigma^*}=M_{\Sigma^*}-M_{\Xi^*}=M_{\Xi^*}-M_\Omega\quad\text{and}\quad M_{\Xi^*}-M_{\Sigma^*}+M_N=\frac{1}{8}(11M_\Lambda-3M_\Sigma). \tag*{(73)} 2 ( M N + M Ξ ) = 3 M Λ + M Σ , M Δ − M Σ ∗ = M Σ ∗ − M Ξ ∗ = M Ξ ∗ − M Ω and M Ξ ∗ − M Σ ∗ + M N = 8 1 ( 11 M Λ − 3 M Σ ) . (73)
The first two are the well-established Gell–Mann-Okubo relations but the third one is specific to H s b H_{\mathrm{sb}} H sb and its treatment in first order perturbation theory. Empirically it is satisfied within a fraction of a percent (1088 MeV versus 1087 MeV). On the other hand, this first order treatment also predicts the ratios ( M Λ − M N ) : ( M Σ − M Λ ) : ( M Ξ − M Σ ) = 1 : 1 : 1 2 (M_{\Lambda}-M_{N}):(M_{\Sigma}-M_{\Lambda}):(M_{\Xi}-M_{\Sigma})=1:1:\frac{1}{2} ( M Λ − M N ) : ( M Σ − M Λ ) : ( M Ξ − M Σ ) = 1 : 1 : 2 1 , while the data have a reversed order: 1 : 0.43 : 0.69 1:0.43:0.69 1 : 0.43 : 0.69 . The inclusion of higher orders improves on that when the product β 2 γ \beta^{2}\gamma β 2 γ is large enough [48 ] . More elegantly, it is possible to diagonalize the Hamiltonian H + H s b H+H_{\mathrm{sb}} H + H sb exactly [49 ] : introducing eight Euler angles via A = A 2 ( α , β , γ ) e − i ν λ 4 A 2 ( α ′ , β ′ , γ ′ ) e − i ρ λ 8 / 3 A=A_{2}(\alpha,\beta,\gamma)e^{-i\nu\lambda_{4}}A_{2}(\alpha^{\prime},\beta^{\prime},\gamma^{\prime})e^{-i\rho\lambda_{8}/\sqrt{3}} A = A 2 ( α , β , γ ) e − i ν λ 4 A 2 ( α ′ , β ′ , γ ′ ) e − i ρ λ 8 / 3 , where A 2 A_{2} A 2 is a two flavor matrix like in Eq. (48) , embedded as in Eq. (65) , and expressing C 2 C_{2} C 2 in terms of differential operators with respect to these angles, the eigenvalue problem is cast into sets of ordinary differential equations with respect to the strangeness changing angle ν \nu ν since 1 − D 88 = 3 2 sin 2 ν 1-D_{88}=\frac{3}{2}\sin^{2}\nu 1 − D 88 = 2 3 sin 2 ν . A typical result for the mass differences is listed in Table 2 . The above mentioned ratios improve to 1 : 0.61 : 0.72 1:0.61:0.72 1 : 0.61 : 0.72 ; though it not a perfect match, the empirical order is reproduced. This is achieved by assuming a small value for the Skyrme parameter e e e , which has the disadvantage that the spin 3 2 \frac{3}{2} 2 3 baryon masses are predicted too low. Yet, refined models also improve on that.
Table 2. Skyrme model prediction for the mass differences with respect to the nucleon using e = 4.0 e=4.0 e = 4.0 [21 ] . For later reference we include the results from the vector meson model (VM) of Section Vector mesons .
Static properties can be computed in parallel to the procedure of the two-flavor model, Section Static properties . But it becomes more involved because additional collective coordinate operators emerge for the symmetry currents. As an example we list the spatial components of the octet axial-vector current (Roman letters run from one to three, except a = 1 , … , 8 a=1,\ldots,8 a = 1 , … , 8 ; Greek letters from four to seven)
A i a = [ a 1 ( r ) δ i k + a 2 ( r ) x ^ i x ^ k ] D a k + [ a 3 ( r ) δ i k + a 4 ( r ) x ^ i x ^ k ] d k α β D a α Ω β + [ a 5 ( r ) δ i k + a 6 ( r ) x ^ i x ^ k ] D a 8 Ω k + symmetry breaking pieces , (74) A^{a}_{i}=\left[a_{1}(r)\delta_{ik}+a_{2}(r)\hat{x}_{i}\hat{x}_{k}\right]D_{ak}+\left[a_{3}(r)\delta_{ik}+a_{4}(r)\hat{x}_{i}\hat{x}_{k}\right]d_{k\alpha\beta}D_{a\alpha}\Omega_{\beta}+\left[a_{5}(r)\delta_{ik}+a_{6}(r)\hat{x}_{i}\hat{x}_{k}\right]D_{a8}\Omega_{k}+\text{symmetry breaking pieces}, \tag*{(74)} A i a = [ a 1 ( r ) δ ik + a 2 ( r ) x ^ i x ^ k ] D ak + [ a 3 ( r ) δ ik + a 4 ( r ) x ^ i x ^ k ] d k α β D a α Ω β + [ a 5 ( r ) δ ik + a 6 ( r ) x ^ i x ^ k ] D a 8 Ω k + symmetry breaking pieces , (74)
where d a b c d_{abc} d ab c are the symmetric S U ( 3 ) SU(3) S U ( 3 ) structure constants. Though tedious, all collective coordinate matrix elements can be computed in the exact diagonalization program mentioned above. The radial functions a 1 ( r ) a_{1}(r) a 1 ( r ) and a 2 ( r ) a_{2}(r) a 2 ( r ) are those from Eq. (55) while the others originate from the Wess-Zumino term. They actually cause a conceptual problem, as they violate ∂ ⋅ A a = 0 \partial\cdot A^{a}=0 ∂ ⋅ A a = 0 in the chiral limit. The reason being that the Wess-Zumino term does not contribute to the classical equations of motion [50 ] . Hence the flavor rotating hedgehog violates PCAC! Induced kaon fields [51 ] are only a partial cure, as they are not orthogonal to the zero mode of the infinitesimal flavor transformation into strangeness direction, i.e., m K > m π m_{K}>m_{\pi} m K > m π is compulsory since otherwise the field equation for the induced components has no solution. Regardless of that problem, together with the vector current pendant, numerous static properties can be computed, for example the hyperon magnetic moments. Results and further references can be found in the review [21 ] . In particular the sensitivity on flavor symmetry breaking can be tested by varying γ \gamma γ which is the crucial entry of the diagonalization program. For example, the axial vector transition matrix elements that change the strangeness of hyperons by one unit have a moderate dependence on γ \gamma γ while the nucleon matrix element of the strange axial current, Δ s = ⟨ N ∣ A i s ∣ N ⟩ \Delta s=\langle N|A^{s}_{i}|N\rangle Δ s = ⟨ N ∣ A i s ∣ N ⟩ , significantly decreases as γ \gamma γ grows. Stated otherwise, the historic flavor symmetry treatment, cf. Ref. [52 ] or chapter IV in Ref. [53 ] , may (to some extent) be a legitimate tool to describe semi-leptonic hyperon decays, but it fails to determine the strangeness content of the proton.
This above described rigid rotor approach assumes that flavor symmetry is still approximately valid. There is the alternative view that it is strongly violated which motivates the so-called bound state approach [54 , 55 ] . Similar to the approach in Section Pion-nucleon scattering , strangeness degrees of freedom, K ( x , t ) K(x,t) K ( x , t ) are incorporated as small amplitude fluctuations about the hedgehog. This can, e.g. be realized by the parameterization
U ( x , t ) = ( A 2 ( t ) 0 0 1 ) ξ H e i Z ξ H ( A 2 † ( t ) 0 0 1 ) with Z = Z ( x , t ) = 2 f K ( 0 K ( x , t ) K † ( x , t ) 0 ) and ξ H 2 = U H . (75) U(x,t)=\left(\begin{array}{cc} A_{2}(t) & 0 \\ 0 & 1 \end{array}\right)\xi_{H}e^{iZ}\xi_{H}\left(\begin{array}{cc} A_{2}^{\dagger}(t) & 0 \\ 0 & 1 \end{array}\right)
\quad\text{with}\quad
Z=Z(x,t)=\frac{\sqrt{2}}{f_{K}}\left(\begin{array}{cc} 0 & K(x,t) \\ K^{\dagger}(x,t) & 0 \end{array}\right)
\quad\text{and}\quad\xi_{H}^{2}=U_{H}. \tag*{(75)} U ( x , t ) = ( A 2 ( t ) 0 0 1 ) ξ H e i Z ξ H ( A 2 † ( t ) 0 0 1 ) with Z = Z ( x , t ) = f K 2 ( 0 K † ( x , t ) K ( x , t ) 0 ) and ξ H 2 = U H . (75)
The kaon fluctuations K ( x , t ) K(x,t) K ( x , t ) are treated to harmonic order. Then the P P P -wave zero mode for rotations into strangeness direction, that emerges for m K = m π m_{K}=m_{\pi} m K = m π , turns into a bound state with energy eigenvalue ω P \omega_{P} ω P when m K > m π m_{K}>m_{\pi} m K > m π . The Wess-Zumino term contributes a term that is linear in the energy eigenvalue and thus lifts the degeneracy of positive and negative strangeness and only the latter has a bound state. Hyperons states are then constructed by (multiple) occupation of this bound state. The quantization of this system, which is discussed at length in the literature, cf. Ref. [55 ] , yields the mass formula
M B = E c l + ∣ S ∣ ω P + 1 2 α 2 [ χ J ( J + 1 ) + ( 1 − χ ) I ( I + 1 ) ] , (76) M_{B}=E_{\mathrm{cl}}+|S|\omega_{P}+\frac{1}{2\alpha^{2}}\left[\chi J(J+1)+(1-\chi)I(I+1)\right], \tag*{(76)} M B = E cl + ∣ S ∣ ω P + 2 α 2 1 [ χ J ( J + 1 ) + ( 1 − χ ) I ( I + 1 ) ] , (76)
for the low-lying P P P -wave baryons with strangeness S = 0 , − 1 , − 2 , − 3 S=0,-1,-2,-3 S = 0 , − 1 , − 2 , − 3 and total spin (isospin) J ( I ) J(I) J ( I ) . The hyperfine splitting parameter χ \chi χ is computed as a spatial integral of the chiral angle and the bound state wave-function. Numerical results for the mass differences are shown in Table 3 .
Table 3. Mass differences of the low-lying baryons with respect to the nucleon in the bound state approach to the Skyrme model with f π = 93 M e V f_\pi=93\,\mathrm{MeV} f π = 93 MeV . The Skyrme model parameter, e = 4.25 e=4.25 e = 4.25 properly reproduces M Δ − M N M_\Delta-M_N M Δ − M N . All data are in MeV.
The hyperon state/wave-function is a combination of products D J 3 , − J 3 ′ ∗ ( Φ , Θ , Ψ ) a 3 † ( ω P ) ∣ 0 ⟩ D^{*}_{J_{3},-J_{3}^{\prime}}(\Phi,\Theta,\Psi)a_{3}^{\dagger}(\omega_{P})|0\rangle D J 3 , − J 3 ′ ∗ ( Φ , Θ , Ψ ) a 3 † ( ω P ) ∣0 ⟩ with the Wigner-D function for the collective rotation, A 2 A_{2} A 2 and a † a^{\dagger} a † is the creation operator for the bound state. The combination goes over J 3 ′ J_{3}^{\prime} J 3 ′ and appropriate Clebsch-Gordan coefficients. The isospin is carried by U H U_{H} U H so that A 2 A_{2} A 2 has zero spin. Hence the rotator must be quantized with integer (half-integer) spin for odd (even) occupation numbers of the bound state. It is then possible to compute static properties like magnetic moments [56 ] . A set of typical results is presented in Table 4 .
Table 4. Predictions for the bound state approach (BSA) for baryon magnetic moments relative to the proton magnetic moment which is predicted too low: 1.78, versus 2.79 (Expt.). Parameters are as in Tab. 3. Also listed are predictions from the rigid rotor approach from the Skyrme model (SK) and the vector meson model (VM) of Section Vector mesons.
We conclude this section with the remark that also the S S S -wave channel contains a bound state solution with negative strangeness. This state has been employed to describe the Λ ( 1405 ) \Lambda(1405) Λ ( 1405 ) resonance [57 ] .
Vector mesons
In Section The Skyrme Model we have seen deficits of the Skyrme model predictions. Some of them can be cured by extending the model such that it contains vector meson fields. Unfortunately matters become very technical quickly. Therefore we will only discuss the concepts here and refer the interested reader to the review articles of Refs. [17 , 18 ] for more details.
The goal is to construct a chiral Lagrangian that contains the vector mesons ω \omega ω and ρ \rho ρ besides the chiral field. These vector mesons have masses m V ≈ 770 M e V m_V \approx770\,\mathrm{MeV} m V ≈ 770 MeV . (We ignore the difference m ω − m ρ ≈ 12 M e V m_\omega-m_\rho\approx12\,\mathrm{MeV} m ω − m ρ ≈ 12 MeV .) Chiral symmetry requires to also consider the fields for their chiral partners, like the axial-vector meson a 1 a_1 a 1 , which is much heavier, m a 1 ≈ 1260 M e V m_{a_1} \approx1260\,\mathrm{MeV} m a 1 ≈ 1260 MeV , and which one would not want to include, at least for simplicity. The starting point is to write a chirally invariant Lagrangian for left- and right-handed vector fields, A L μ A_L^\mu A L μ and A R μ A_R^\mu A R μ , which are sums and differences of vector and axial-vector meson fields. Among standard terms like tr [ F L , μ ν F L μ ν ] \operatorname{tr}\left[F_{L,\mu\nu}F_L^{\mu\nu}\right] tr [ F L , μν F L μν ] this Lagrangian contains unconventional, chirally invariant terms like [29 ]
tr [ F L , μ ν U F R μ ν U † ] , tr [ A L , μ U A R μ U † ] or ε μ ν ρ σ tr [ ∂ μ A L ν ∂ ρ U A R σ U † ] . (77) \operatorname{tr}\left[F_{L,\mu\nu}UF_R^{\mu\nu}U^\dagger\right],\quad\operatorname{tr}\left[A_{L,\mu}UA_R^\mu U^\dagger\right]\quad\text{or}\quad\varepsilon^{\mu\nu\rho\sigma}\operatorname{tr}\left[\partial^\mu A_L^\nu\partial^\rho U A_R^\sigma U^\dagger\right]. \tag*{(77)} tr [ F L , μν U F R μν U † ] , tr [ A L , μ U A R μ U † ] or ε μν ρ σ tr [ ∂ μ A L ν ∂ ρ U A R σ U † ] . (77)
To eliminate the a 1 a_1 a 1 field without violating chiral symmetry the vector mesons are parameterized as
A L μ = ξ ( V μ + i g ∂ μ ) ξ † , A R μ = ξ † ( V μ + i g ∂ μ ) ξ and U = ξ 2 1 2 × 2 . (78) A_L^\mu=\xi\left(V^\mu+\frac{i}{g}\partial^\mu\right)\xi^\dagger,\quad A_R^\mu=\xi^\dagger\left(V^\mu+\frac{i}{g}\partial^\mu\right)\xi\quad\text{and}\quad U=\xi^2\mathbf{1}_{2\times2}. \tag*{(78)} A L μ = ξ ( V μ + g i ∂ μ ) ξ † , A R μ = ξ † ( V μ + g i ∂ μ ) ξ and U = ξ 2 1 2 × 2 . (78)
The ω \omega ω and ρ \rho ρ fields are then identified as the isoscalar and isovector components of the new field, V μ = ω μ 1 2 × 2 + ρ μ ⋅ τ V^\mu=\omega^\mu\mathbf{1}_{2\times2}+\boldsymbol{\rho}^{\mu}\cdot\boldsymbol{\tau} V μ = ω μ 1 2 × 2 + ρ μ ⋅ τ . Possible model parameters, i.e., the coupling constant g g g and the coefficients of the terms in Eq. (77) and others, are determined from vector meson properties like their masses and the decay widths for ρ → π π \rho\to\pi\pi ρ → π π or ω → π π π \omega\to\pi\pi\pi ω → π π π [58 ] . Stabilizing terms, like the Skyrme term or L 6 \mathcal{L}_6 L 6 discussed in Section The Skyrme model are not included. On the contrary, the local approximation in which m V 2 V μ ≫ ∂ 2 V μ m_V^2V_\mu\gg\partial^2V_\mu m V 2 V μ ≫ ∂ 2 V μ yields F μ ν ( ρ ) ∼ [ α μ , α ν ] F_{\mu\nu}(\rho)\sim[\alpha_\mu,\alpha_\nu] F μν ( ρ ) ∼ [ α μ , α ν ] and ω μ ∼ B μ \omega_\mu\sim B_\mu ω μ ∼ B μ . That is, in this limit the vector meson coupling generates those stabilizing terms, but now with coefficients known from meson properties. Of course, the ongoing calculations will not adopt the local approximation. The so-called hidden symmetry approach [59 ] obtains the same model Lagrangian by writing U = ξ L † ξ R U=\xi_L^\dagger\xi_R U = ξ L † ξ R and elevating the transformation ξ L → h ξ L L \xi_L\to h\xi_L L ξ L → h ξ L L , ξ R → h ξ R R \xi_R\to h\xi_R R ξ R → h ξ R R to a local symmetry with h = h ( x ) h=h(x) h = h ( x ) by means of introducing V μ V_\mu V μ as a gauge field.
The spherically symmetric soliton configuration has three profile functions F ( r ) F(r) F ( r ) , ω ( r ) \omega(r) ω ( r ) and G ( r ) G(r) G ( r ) in
ξ H = exp [ i 2 x ^ ⋅ τ F ( r ) ] , ω 0 = ω ( r ) g and ρ i = G ( r ) g r ( x × τ ) i . (79) \xi_H=\exp\left[\frac{i}{2}\hat{x}\cdot\boldsymbol{\tau}F(r)\right],\quad\omega^0=\frac{\omega(r)}{g}\quad\text{and}\quad\rho^i=\frac{G(r)}{gr}(\boldsymbol{x}\times\boldsymbol{\tau})_i. \tag*{(79)} ξ H = exp [ 2 i x ^ ⋅ τ F ( r ) ] , ω 0 = g ω ( r ) and ρ i = g r G ( r ) ( x × τ ) i . (79)
The space components of ω μ \omega^\mu ω μ and the time components ρ 0 \rho^0 ρ 0 are zero. A set of profiles that solves the classical equations is displayed in Figure 4 . The result G ( 0 ) = − 2 G(0)=-2 G ( 0 ) = − 2 is not a singularity because the coupling to the chiral angle is ( G + 1 − cos F ) 2 (G+1-\cos F)^2 ( G + 1 − cos F ) 2 [58 ] .
Figure 4. Profile functions for the vector meson soliton. The vector meson profile ω \omega ω is measured in units of the vector meson mass m V m_V m V .
A major difference to the Skyrme model is that even for two flavors the quantization is not as straightforward as in Section Quantization. On top of the time-dependent rotations, Eq. (39) , field components that vanish classically get induced [60 , 61 ] . These components are incorporated by writing
ξ = exp [ i 2 f π η ( r ) x ⋅ Ω ] A ( t ) ξ H A † ( t ) , ρ 0 = 1 2 g A ( t ) [ ξ 1 ( r ) Ω + ξ 2 ( r ) ( x ⋅ Ω ) x ] A † ( t ) and ω i = ϕ ( r ) 2 g ε i j k Ω j x k . (80) \xi=\exp\left[\frac{i}{2f_\pi}\eta(r)\boldsymbol{x}\cdot\boldsymbol{\Omega}\right]A(t)\xi_H A^\dagger(t),\quad\rho^0=\frac{1}{2g}A(t)\left[\xi_1(r)\boldsymbol{\Omega}+\xi_2(r)(\boldsymbol{x}\cdot\boldsymbol{\Omega})\boldsymbol{x}\right]A^\dagger(t)\quad\text{and}\quad\omega^i=\frac{\phi(r)}{2g}\varepsilon_{ijk}\Omega_jx_k. \tag*{(80)} ξ = exp [ 2 f π i η ( r ) x ⋅ Ω ] A ( t ) ξ H A † ( t ) , ρ 0 = 2 g 1 A ( t ) [ ξ 1 ( r ) Ω + ξ 2 ( r ) ( x ⋅ Ω ) x ] A † ( t ) and ω i = 2 g ϕ ( r ) ε ij k Ω j x k . (80)
The four profile functions η \eta η , ξ 1 , 2 \xi_{1,2} ξ 1 , 2 and ϕ \phi ϕ obey inhomogeneous linear differential equations that arise from extremizing the moment of inertia, α 2 \alpha^2 α 2 . The classical soliton provides the source terms in these equations. Typical solutions are shown in Figure 5 . The calculations of baryon properties are quite technical and we merely summarize some of the substantial improvements compared to the Skyrme model:
Figure 5. Profile functions for the induced components of the vector meson soliton.
(Approximate) vector meson dominance relates the ω \omega ω field to the iso-scalar density, V 0 0 V^0_0 V 0 0 which is thus subject to an equation of the form,
( ∂ 2 − m V 2 ) V 0 0 ∼ − m V 2 B 0 ( r ) . (\partial^2-m_V^2)V^0_0\sim-m_V^2B_0(r). ( ∂ 2 − m V 2 ) V 0 0 ∼ − m V 2 B 0 ( r ) . Multiplication by r 2 r^2 r 2 and integration over space yields [60 ] 10 ^{10} 10
⟨ r 2 ⟩ V 0 0 = ⟨ r 2 ⟩ B + 6 m V 2 , \langle r^2\rangle_{V^0_0}=\langle r^2\rangle_B+\frac{6}{m_V^2}, ⟨ r 2 ⟩ V 0 0 = ⟨ r 2 ⟩ B + m V 2 6 , where ⟨ r 2 ⟩ B \langle r^2\rangle_B ⟨ r 2 ⟩ B is the radius of the baryon density and equals the isoscalar radius in the Skyrme model. The additional 0.4fm bring the too small predictions in Tab. 1 into better agreement with data.
The η \eta η profile is induced via its coupling to the vector meson fields. Though it is small, cf. Figure 5 , it yields the important result
tr [ τ 3 ( U − U † ) ] ∼ D 3 i x ^ i ( x ^ ⋅ Ω ) η ( r ) sin F ( r ) , 0. \operatorname{tr}\left[\tau_3\left(U-U^\dagger\right)\right]\sim D_{3i}\hat{x}_i(\hat{x}\cdot\Omega)\eta(r)\sin F(r),\quad0. tr [ τ 3 ( U − U † ) ] ∼ D 3 i x ^ i ( x ^ ⋅ Ω ) η ( r ) sin F ( r ) , 0. As a consequence the two flavor vector meson model, in contrast to the Skyrme model, predicts a non-zero strong interaction contribution to the neutron-proton mass difference: ( m n − m p ) s t r . ≈ 1.2 M e V (m_n-m_p)_{\mathrm{str.}}\approx1.2\,\mathrm{MeV} ( m n − m p ) str. ≈ 1.2 MeV [64 ] .
It was already mentioned that the vector mesons replace higher order stabilizing terms in the Skyrme model. In the context of the adiabatic approach to meson-baryon scattering the dynamical vector mesons replace contact interactions by a propagator (like the transition from the Fermi model to the Standard Model of particle physics with massive vector mesons):
G 2 m V 2 → G 2 m V 2 − k 2 . \frac{G^2}{m_V^2}\to\frac{G^2}{m_V^2-k^2}. m V 2 G 2 → m V 2 − k 2 G 2 . In turn this leads to M i j 00 = M i j r r M^{00}_{ij}=M^{rr}_{ij} M ij 00 = M ij r r and solves the ever-rising phase shift problem [18 ] . We also refer to that article for detailed comparison with data and the application of the distorted wave-functions that emerge from this scattering problem to meson-photon production processes.
The S U ( 3 ) SU(3) S U ( 3 ) vector meson model calculations are as cumbersome as the one for the scattering in d) [65 ] . It brings into the game further induced components as in Eq. (80) but now related to the angular velocity Ω 4 , … , Ω 7 \Omega_4,\ldots,\Omega_7 Ω 4 , … , Ω 7 . These components are obtained by extremizing the second moment of inertia β 2 \beta^2 β 2 . Moreover, this model comes with additional flavor symmetry breaking structures like
∑ i = 1 3 D 8 i J i , ∑ α = 1 7 D 8 α R α , ∑ i = 1 3 D 8 i D 8 i and ∑ α = 1 7 D 8 α D 8 α . (81) \sum_{i=1}^{3}D_{8i}J_i,\qquad\sum_{\alpha=1}^{7}D_{8\alpha}R_\alpha,\qquad\sum_{i=1}^{3}D_{8i}D_{8i}\quad\text{and}\quad\sum_{\alpha=1}^{7}D_{8\alpha}D_{8\alpha}. \tag*{(81)} i = 1 ∑ 3 D 8 i J i , α = 1 ∑ 7 D 8 α R α , i = 1 ∑ 3 D 8 i D 8 i and α = 1 ∑ 7 D 8 α D 8 α . (81)
In particular the first of these operators is very welcome to improve on the predictions for the ratios ( M Λ − M N ) : ( M Σ − M Λ ) : ( M Ξ − M Σ ) (M_\Lambda-M_N):(M_\Sigma-M_\Lambda):(M_\Xi-M_\Sigma) ( M Λ − M N ) : ( M Σ − M Λ ) : ( M Ξ − M Σ ) .
Some of the many prediction of this model have already been listed in Tabs. 2 and 4. For additional ones we refer to the many references quoted in the review articles [17 , 21 ] on top of those listed under the above items.
Quarks and Skyrmions
So far we have considered chiral solitons that were directly written as functions of the chiral field U U U . However, there is an alternative, and essentially equivalent approach that starts from a chirally invariant formulation of the quark flavor dynamics. The number of publications on this topic is legion and thus we just mention the two main review articles [66 , 67 ] . Most of the many publications repeat the studies under items a), b), c) and e) above. There are improvements when it comes to agreement with empirical data, but there is not much of a conceptual difference. An exception may be the incorporation of sub-leading 1 / N C 1/N_C 1/ N C terms for the axial vector charge, g A g_A g A and the iso-vector magnetic moment, μ p − μ n \mu_p-\mu_n μ p − μ n [68 ] , though there are inconsistencies with PCAC in these approaches [69 ] . In any event, we will focus here on a particular application of quark models that (at least without further approximations) is not accessible in purely meson theories: nucleon structure functions.
It has been long known [70 , 71 , 72 ] that when the Dirac equation ( i ̸ ∂ − m ) q = 0 (i\not\!\partial-m)q=0 ( i ∂ − m ) q = 0 is augmented with a chiral interaction of the form q ˉ U H q \bar{q}U_Hq q ˉ U H q , a strongly bound quark valence quark solution emerges with energy eigenvalue ∣ ϵ v ∣ < m |\epsilon_v|<m ∣ ϵ v ∣ < m for wide enough chiral angles, F ( r ) F(r) F ( r ) . However, considering only this single level is merely the quasi-classical approximation [70 ] which does not have a well defined expansion parameter. Rather, one needs to consider the semi-classical approximation which includes the contribution from the distorted Dirac sea: − 1 2 ∑ n [ ∣ ϵ n ∣ − ∣ ϵ n ( 0 ) ∣ ] -\frac{1}{2}\sum_n[|\epsilon_n|-|\epsilon_n^{(0)}|] − 2 1 ∑ n [ ∣ ϵ n ∣ − ∣ ϵ n ( 0 ) ∣ ] . Here ϵ n \epsilon_n ϵ n and ϵ n ( 0 ) \epsilon_n^{(0)} ϵ n ( 0 ) are quark energy eigenvalues with and without coupling to U H U_H U H , respectively. The leading order of the semi-classical expansion is determined from the one-loop effective action
A F = − i Log Det Λ [ i ̸ ∂ − m ( U H ) γ 5 ] with ( U H ) γ 5 = exp [ i x ^ ⋅ τ γ 5 F ( r ) ] . (82) \mathcal{A}_F=-i\operatorname{Log}\operatorname{Det}_{\Lambda}\left[i\not\!\partial-m(U_H)^{\gamma_5}\right]\qquad\text{with}\qquad(U_H)^{\gamma_5}=\exp\left[i\hat{x}\cdot\boldsymbol{\tau}\gamma_5F(r)\right]. \tag*{(82)} A F = − i Log Det Λ [ i ∂ − m ( U H ) γ 5 ] with ( U H ) γ 5 = exp [ i x ^ ⋅ τ γ 5 F ( r ) ] . (82)
This part of the action can be motivated from QCD by a number of successive steps: (i) approximate the gluon exchange by an effective four quark interaction 1 4 ( q ˉ q ) 2 \frac{1}{4}(\bar{q}q)^2 4 1 ( q ˉ q ) 2 , (ii) provide an auxiliary meson field Φ \Phi Φ to write this interaction as q ˉ Φ q − Φ 2 \bar{q}\Phi q-\Phi^2 q ˉ Φ q − Φ 2 , (iii) only quark bilinears are left and the functional integral can be performed, (iv) for sufficiently large coupling a non-zero VEV, ⟨ Φ ⟩ = m \langle\Phi\rangle=m ⟨ Φ ⟩ = m emerges dynamically as the constituent quark mass, (v) introduce fluctuations about the VEV in form of the chiral field via Φ = m U \Phi=mU Φ = m U , (vi) the fluctuating modes are the pseudoscalar mesons, Eq. (23) [73 ] . These steps are rather sketchy and the full approach is quite sophisticated. In particular it is crucial to maintain chiral symmetry throughout the calculation. Unfortunately, confinement (if it ever was there) and renormalizability are lost. The latter is indicated by the cut-off Λ \Lambda Λ in Eq. (82) which acquires a physical meaning in the model. In fact, it is related to the pion mass and decay constant. In many applications Schwinger’s proper-time regularization scheme [74 ] has been used. Here we will, however, consider a variant of the Pauli-Villars scheme, which, as an important feature is implemented on the level of Eq. (82) . The final classical energy functional then reads
E t o t [ F ] = N C 2 [ 1 + sign ( ϵ v ) ] ϵ v − N C 2 ∑ n { ∣ ϵ n ∣ − ϵ n 2 + Λ 2 + 1 2 ∣ ϵ n ∣ ϵ n 2 + Λ 2 − ( ϵ n → ϵ n ( 0 ) ) } + m π 2 f π 2 ∫ d 3 r [ 1 − cos F ] . (83) E_{\mathrm{tot}}[F]=\frac{N_C}{2}\left[1+\operatorname{sign}(\epsilon_v)\right]\epsilon_v-\frac{N_C}{2}\sum_n\left\{|\epsilon_n|-\sqrt{\epsilon_n^2+\Lambda^2}+\frac{1}{2}\frac{|\epsilon_n|}{\sqrt{\epsilon_n^2+\Lambda^2}}-\left(\epsilon_n\rightarrow\epsilon_n^{(0)}\right)\right\}+m_\pi^2f_\pi^2\int d^3r\,[1-\cos F]. \tag*{(83)} E tot [ F ] = 2 N C [ 1 + sign ( ϵ v ) ] ϵ v − 2 N C n ∑ { ∣ ϵ n ∣ − ϵ n 2 + Λ 2 + 2 1 ϵ n 2 + Λ 2 ∣ ϵ n ∣ − ( ϵ n → ϵ n ( 0 ) ) } + m π 2 f π 2 ∫ d 3 r [ 1 − cos F ] . (83)
where the ϵ n \epsilon_n ϵ n are the eigenvalues of Dirac Hamiltonian h = α ⋅ p + β m ( U H ) γ 5 h=\boldsymbol{\alpha}\cdot\boldsymbol{p}+\beta m(U_H)^{\gamma_5} h = α ⋅ p + β m ( U H ) γ 5 . Hence these eigenvalues are functionals of U H U_H U H and thus of the chiral angle F ( r ) F(r) F ( r ) . The factor N C 2 [ 1 + sign ( ϵ v ) ] \frac{N_C}{2}[1+\operatorname{sign}(\epsilon_v)] 2 N C [ 1 + sign ( ϵ v )] emerges in the first term because the Dirac sea carries baryon number when ϵ v < 0 \epsilon_v<0 ϵ v < 0 and the bound valence level should not be occupied explicitly. The chiral angle is obtained by self-consistently solving this eigenvalue problem and minimising E t o t [ F ] E_{\mathrm{tot}}[F] E tot [ F ] . Spin and isospin states are generated by the collective coordinate formalism as in Section Quantization. Substituting the chiral field from Eq. (39) yields the modified Dirac Hamiltonian h Ω = A ( α ⋅ p + 1 2 τ ⋅ Ω + β m ( U H ) γ 5 ) A † h_\Omega=A(\boldsymbol{\alpha}\cdot\boldsymbol{p}+\frac{1}{2}\boldsymbol{\tau}\cdot\boldsymbol{\Omega}+\beta m(U_H)^{\gamma_5})A^\dagger h Ω = A ( α ⋅ p + 2 1 τ ⋅ Ω + β m ( U H ) γ 5 ) A † . After changing the integration variables to q = A q ~ q=A\tilde{q} q = A q ~ , the action, Eq. (82) , is then expanded to quadratic order in the angular velocities yielding a Lagrange function of same form as in Eq. (50) . Here the moment of inertia, α 2 \alpha^2 α 2 is obtained as a regularized double sum over energy eigenvalues ϵ n \epsilon_n ϵ n and ϵ m \epsilon_m ϵ m . As mentioned, there are relations between the model parameters and pion observables. Eventually only the constituent quark mass remains the only free parameter which, similar to the Skyrme parameter, is typically tuned to reproduce 3 2 α 2 = M Δ − M N ≈ 293 M e V \frac{3}{2\alpha^2}=M_\Delta-M_N\approx293\,\mathrm{MeV} 2 α 2 3 = M Δ − M N ≈ 293 MeV .
Rather than considering static baryon properties in chiral quark soliton models we immediately focus on nucleon structure functions. There have been several explorations that compute the structure functions by applying expressions for QCD distributions11 to the eigenstates of h h h (or h Ω h_\Omega h Ω ), i.e., the model (constituent) quarks [75 , 76 , 77 , 78 , 79 ] . This seems as an over-interpretation of the model. Rather we only want to identify the symmetry currents from QCD, not the quark fields themselves. Furthermore, when starting from Eq. (82) the subtle issue of how to regularize the vacuum contribution to the structure functions is stipulated from the onset.
The interaction vertex for the disintegration of the nucleon is the matrix element of the (electromagnetic) current J μ ( ξ ) J_\mu(\xi) J μ ( ξ ) . The cross-section contains the squared absolute value of this matrix element and we sum/integrate over all final states subject to energy momentum conservation. This defines the hadron tensor for deep-inelastic electron nucleon scattering
W μ ν ( p , q ; s ) = 1 4 π ∑ X ⟨ p , s ∣ J μ ( 0 ) ∣ X ⟩ ⟨ X ∣ J ν † ( 0 ) ∣ p , s ⟩ ( 2 π ) 4 δ 4 ( p + q − p X ) = 1 4 π ∫ d 4 ξ e i q ⋅ ξ ⟨ p , s ∣ [ J μ ( ξ ) , J ν † ( 0 ) ] ∣ p , s ⟩ . (84) W_{\mu\nu}(p,q;s)=\frac{1}{4\pi}\sum_X\langle p,s|J_\mu(0)|X\rangle\langle X|J_\nu^\dagger(0)|p,s\rangle(2\pi)^4\delta^4(p+q-p_X)=\frac{1}{4\pi}\int d^4\xi\,e^{iq\cdot\xi}\langle p,s|[J_\mu(\xi),J_\nu^\dagger(0)]|p,s\rangle. \tag*{(84)} W μν ( p , q ; s ) = 4 π 1 X ∑ ⟨ p , s ∣ J μ ( 0 ) ∣ X ⟩ ⟨ X ∣ J ν † ( 0 ) ∣ p , s ⟩ ( 2 π ) 4 δ 4 ( p + q − p X ) = 4 π 1 ∫ d 4 ξ e i q ⋅ ξ ⟨ p , s ∣ [ J μ ( ξ ) , J ν † ( 0 )] ∣ p , s ⟩ . (84)
where p p p and s s s the nucleon momentum and spin. The momentum of the virtual photon is the difference of the momenta of the initial and final nucleons: q = k − k ′ q=k-k' q = k − k ′ with q 0 > 0 q_0>0 q 0 > 0 and Q 2 = − q 2 > 0 Q^2=-q^2>0 Q 2 = − q 2 > 0 since the interaction is inelastic and space-like. The second part in Eq. (84) results from translational invariance. The tensor W μ ν W_{\mu\nu} W μν defines Lorentz invariant form factors via the decomposition
W μ ν ( p , q ; s ) = ( − g μ ν + q μ q ν q 2 ) M N W 1 ( x , Q 2 ) + ( p μ − q μ p ⋅ q q 2 ) ( p ν − q ν p ⋅ q q 2 ) 1 M N W 2 ( x , Q 2 ) + i ϵ μ ν λ σ q λ M N p ⋅ q ( [ G 1 ( x , Q 2 ) + G 2 ( x , Q 2 ) ] s σ − q ⋅ s q ⋅ p p σ G 2 ( x , Q 2 ) ) . (85) W_{\mu\nu}(p,q;s)=\left(-g_{\mu\nu}+\frac{q_\mu q_\nu}{q^2}\right)M_NW_1(x,Q^2)+\left(p_\mu-q_\mu\frac{p\cdot q}{q^2}\right)\left(p_\nu-q_\nu\frac{p\cdot q}{q^2}\right)\frac{1}{M_N}W_2(x,Q^2)+i\epsilon_{\mu\nu\lambda\sigma}\frac{q^\lambda M_N}{p\cdot q}\left([G_1(x,Q^2)+G_2(x,Q^2)]s^\sigma-\frac{q\cdot s}{q\cdot p}p^\sigma G_2(x,Q^2)\right). \tag*{(85)} W μν ( p , q ; s ) = ( − g μν + q 2 q μ q ν ) M N W 1 ( x , Q 2 ) + ( p μ − q μ q 2 p ⋅ q ) ( p ν − q ν q 2 p ⋅ q ) M N 1 W 2 ( x , Q 2 ) + i ϵ μν λσ p ⋅ q q λ M N ( [ G 1 ( x , Q 2 ) + G 2 ( x , Q 2 )] s σ − q ⋅ p q ⋅ s p σ G 2 ( x , Q 2 ) ) . (85)
with the kinematical variables x = Q 2 2 M N ν x=\frac{Q^2}{2M_N\nu} x = 2 M N ν Q 2 and ν = p ⋅ q M N \nu=\frac{p\cdot q}{M_N} ν = M N p ⋅ q . The structure functions are defined as the limit Q 2 → ∞ Q^2\to\infty Q 2 → ∞ of the above form factors at fixed x x x . This limit is often called the Bjorken limit and x x x is the Bjorken variable.
The key ingredient for the model calculation is the relation between the hadron tensor and the (forward virtual) Compton amplitude, T μ ν a b T_{\mu\nu}^{ab} T μν ab
W μ ν ( p , q ; s ) = 1 2 π Im T μ ν ( p , q ; s ) with T μ ν ( p , q ; s ) = i ∫ d 4 ξ e i q ⋅ ξ ⟨ p , q ; s ∣ T { J μ ( ξ ) J ν ( 0 ) } ∣ p , q ; s ⟩ . (86) W_{\mu\nu}(p,q;s)=\frac{1}{2\pi}\operatorname{Im}T_{\mu\nu}(p,q;s)
\qquad\text{with}\qquad
T_{\mu\nu}(p,q;s)=i\int d^{4}\xi\,e^{iq\cdot\xi}\langle p,q;s\vert T\{J_{\mu}(\xi)J_{\nu}(0)\}\vert p,q;s\rangle. \tag*{(86)} W μν ( p , q ; s ) = 2 π 1 Im T μν ( p , q ; s ) with T μν ( p , q ; s ) = i ∫ d 4 ξ e i q ⋅ ξ ⟨ p , q ; s ∣ T { J μ ( ξ ) J ν ( 0 )} ∣ p , q ; s ⟩ . (86)
This time-ordered product in T μ ν T_{\mu\nu} T μν is obtained from first principles via the functional derivative
T { J μ ( ξ ) J ν ( 0 ) } = δ 2 δ v μ ( ξ ) δ v ν ( 0 ) A F ( v ) ∣ v μ = 0 , (87) T\{J_{\mu}(\xi)J_{\nu}(0)\}=\left.\frac{\delta^{2}}{\delta v^{\mu}(\xi)\delta v^{\nu}(0)}\mathcal{A}_{\mathrm{F}}(v)\right|_{v_{\mu}=0}, \tag*{(87)} T { J μ ( ξ ) J ν ( 0 )} = δ v μ ( ξ ) δ v ν ( 0 ) δ 2 A F ( v ) v μ = 0 , (87)
where v μ v_{\mu} v μ is the photon field introduced by minimal substitution in Eq. (82) to implement electromagnetic gauge invariance. A main simplification arising from the Bjorken limit in conjunction with the Pauli-Villars scheme is that one of the two propagators ( i ∂ t − h ) − 1 (i\partial_{t}-h)^{-1} ( i ∂ t − h ) − 1 can be taken to be the one of a free, massless fermion, (i\beta\slashed{\partial})^{-1} . This is indicated in Figure 6 . Whether this simplification holds in other regularization schemes is unclear.
Figure 6. Two photon coupling to fermion loop. Thick lines are the full fermion propagators in the soliton background. The thin line in the loop on the right represents a free (massless) fermion propagator, (i\beta\slashed{\partial})^{-1} . Dashed lines denote Cutkosky cuts to extract the imaginary part, Eq. (86) .
Now the stage is set to evaluate Eqs. (86) and (87) within this chiral soliton model. This results in bulky expressions [80 , 81 ] for the structure functions that will not even be indicated here. The numerical simulation yields the structure functions in the nucleon rest frame (RF) at a low renormalization scale μ 2 \mu^{2} μ 2 . In the rest frame the calculated structure functions have support outside the kinematically allowed regime x ∈ [ 0 , 1 ] x\in[0,1] x ∈ [ 0 , 1 ] because the soliton is localized. This is cured by a Lorentz boost to the infinite momentum frame (IMF) [82 , 83 ] , which also makes contact with the parton model picture of structure functions. The low renormalization scale is a new parameter in the model that sets the initial condition for the DGLAP-evolution program [84 , 85 , 86 ] .12 As an example we compare the model prediction for the axial structure function g 1 g_{1} g 1 to data in Figure 7 . Proton data are directly available, while neutron data are accessible via scattering electrons off 3 H e {}^{3}\mathrm{He} 3 He .
Figure 7. Model prediction for the longitudinal polarized proton structure functions. Left panel: g 1 p ( x ) g_1^p(x) g 1 p ( x ) ; right panel: g 1 3 H e ( x ) g_1^{3\mathrm{He}}(x) g 1 3 He ( x ) . These functions are ”DGLAP” evolved from μ 2 = 0.4 G e V 2 \mu^2 = 0.4\,\mathrm{GeV}^2 μ 2 = 0.4 GeV 2 to Q 2 = 3 G e V 2 Q^2 = 3\,\mathrm{GeV}^2 Q 2 = 3 GeV 2 after being projected from the RF to the IMF. Data are from Refs. [88 , 89 ] for the proton and from Ref. [90 ] for helium. In the latter case E E E refers to the energy of the incident electron.
Though there are discrepancies, we nevertheless observe that chiral soliton models reproduce the main characteristics of the data and that these models indeed have their say on nucleon structure functions.
Further applications
There are many other applications of chiral solitons in hadron physics and beyond. We list some of them with going into much detail.
Flavor symmetry breaking and soliton extension
According to Eq. (32) the asymptotic behavior of the flavor rotating soliton, Eq. (66) is governed by the pion mass. However, the kaon mass should enter for the strangeness degrees of freedom so that a hedgehog pointing into kaon direction would be narrower. There are two approaches that address this problem. In the slow-rotator approach the strangeness changing angle ν \nu ν defined after Eq. (7) is considered a constant parameter in the classical field equation that arises from the Euler-Lagrange equations for L S K + L s b \mathcal{L}_{\mathrm{SK}}+\mathcal{L}_{\mathrm{sb}} L SK + L sb ; recall that L s b \mathcal{L}_{\mathrm{sb}} L sb contains 1 − D 88 ( A ) = 3 2 sin 2 ν 1-D_{88}(A)=\frac{3}{2}\sin^{2}\nu 1 − D 88 ( A ) = 2 3 sin 2 ν . In turn the chiral angle depends parametrically on ν \nu ν and so do the classical mass and the moments of inertia. This parameter dependence is included in the energy eigenvalue problem formulated as differential equations with respect to ν \nu ν [91 ] . In the other, the breathing mode approach, the extension of the soliton is quantized as a collective coordinate. Here the starting point is the field parameterization13
U ( t , x ) = A ( t ) U H ( s ( t ) x ) A † ( t ) . (88) U(t,x)=A(t)U_{\mathrm{H}}(s(t)x)A^{\dagger}(t). \tag*{(88)} U ( t , x ) = A ( t ) U H ( s ( t ) x ) A † ( t ) . (88)
In Section Extension to SU3 we have discussed the leading order treatment [45 ] of flavor symmetry breaking when quantizing the collective rotations A ( t ) A(t) A ( t ) but then immediately turned to the exact diagonalization [49 ] . The results of the latter can essentially be reproduced in perturbation theory when expanding up to third order [48 ] . In doing so, the spin 1 2 \frac{1}{2} 2 1 baryons are no longer pure octet states but acquire admixtures from spin 1 2 \frac{1}{2} 2 1 states with the same flavor quantum numbers in higher dimensional S U ( 3 ) SU(3) S U ( 3 ) representations like the anti-decuplet and the 27 t h 27^{\mathrm{th}} 2 7 th -plet. With the ansatz of Eq. (88) we get an attractive potential for the scaling variable s s s that is proportional to D 88 ( A ) D_{88}(A) D 88 ( A ) , i.e., the strength of the attraction depends on the considered element in a particular S U ( 3 ) SU(3) S U ( 3 ) representation. When quantizing s ( t ) s(t) s ( t ) the eigenstates are thus combinations of elements of different S U ( 3 ) SU(3) S U ( 3 ) representations and radial excitations. Though the nucleon still has its dominant contribution from the radial ground state in the octet, the Roper (1440) resonance has major contributions from the first radial excitation. in the octet and the nucleon radial ground states in the anti-decuplet (called N ′ N' N ′ in Figure 8 ) and the 27th-plet [93 ] . In either case, there are substantial improvements for the predictions of baryon properties in the two approaches. As an example we briefly discuss the magnetic moments of the spin 1 2 \frac{1}{2} 2 1 baryons. From Table 4 we see that the rigid rotator results closely follow the U U U -spin relations: μ p = μ Σ + \mu_p = \mu_{\Sigma^+} μ p = μ Σ + (1 : 0.87), μ n = μ Ξ 0 = 2 μ Λ \mu_n = \mu_{\Xi^0} = 2\mu_\Lambda μ n = μ Ξ 0 = 2 μ Λ (−0.68 : −0.45 : −0.44) and μ Σ − = μ Ξ − \mu_{\Sigma^-} = \mu_{\Xi^-} μ Σ − = μ Ξ − (−0.42 : −0.25). The model results substantially deviate from the experimental data in parenthesis (in units of μ p \mu_p μ p ). The slow-rotator improves these relations to (1 : 0.85), (−0.83 : −0.54 : −0.50) and (−0.40 : −0.20) and so does the breathing mode approach: (1 : 0.78), (−0.89 : −0.41 : −0.39) and (−0.47 : −0.14); which even overshoots the empirical ratios. Including a scalar field as motivated by the QCD scale anomaly [94 ] has a mitigating effect on this overshooting as it does for the mass differences [95 ] . Either of the two approaches corroborates that the extension of the soliton should reflect flavor symmetry breaking.
Figure 8. The flavor content of the anti-decuplet. The quantum numbers are isospin projection (I 3 I_3 I 3 ) and hypercharge (Y Y Y ) which is the sum of baryon-and strangeness number. Note that the strange quark has strangeness negative one.
Pentaquarks
Chiral soliton models had another revival around the beginning of the millennium in the context of the (in)famous pentaquark, Θ + \Theta^+ Θ + . The relevance of higher dimensional S U ( 3 ) SU(3) S U ( 3 ) representations has already been mentioned in the context of flavor symmetry breaking. The lowest dimensional such representation is the anti-decuplet whose particle content is displayed in Figure 8 .
Equally interesting is the fact that these representations contain exotic baryon states with quantum numbers that cannot be described as three quark composites. The most prominent exotic baryon is the Θ + \Theta^+ Θ + whose quantum numbers can only be represented by a quark composition of two up, two down and an anti-strange quark. The relevance of these, eventually low-mass, exotic states in chiral soliton models was recognized quite early [96 , 97 , 98 ] . Yet, it only attracted appreciable recognition when it was conjectured that in particular the Θ + \Theta^+ Θ + was long-lived [99 ] 14 and when there were experimental indications that this was indeed the case [101 ] . In the meantime refined experiments have suggested otherwise unless the width of the Θ + \Theta^+ Θ + is two orders of magnitude smaller than a typical one of a hadron resonance. The current situation has been recently sketched in Ref. [102 ] . Yet, there are reservations on the width estimate of Refs. [99 , 103 ] . The approach is based on determining a Yukawa coupling between exotic and ground state baryons as well as pseudoscalar mesons from the model axial vector current, Eq. (74) . First, the emergence of a Yukawa coupling, i.e., a term linear in the fluctuations, in soliton models is an artifact (shortcoming?) of not knowing the exact time-dependent solution but rather approximating it by Eq. (66) . Second, this axial vector current detour employs the Goldberger-Treiman relation [104 ] , which assumes that the Yukawa coupling, as a function of momentum transfer, does not (significantly) vary between zero and the meson mass. This may not be the case for m K ≃ 495 M e V m_K \simeq495\,\mathrm{MeV} m K ≃ 495 MeV . Furthermore this relation utilizes PCAC which, as mentioned before does not hold for the current in Eq. (74) . To avoid these issues, Ref. [105 ] considers kaon-nucleon scattering similar to the approach in Section Pion-nucleon scattering but beyond the adiabatic approximation and extracts the resonance parameters from scattering data (as experiment does). In that approach the Θ + \Theta^+ Θ + is predicted to be about 700 M e V 700\,\mathrm{MeV} 700 MeV heavier than the nucleon (which is not low) with a width of something like 40 M e V 40\,\mathrm{MeV} 40 MeV which is only slightly smaller than a typical hadron width.
Heavy flavor symmetry
For mesons containing a single charm or bottom quark the pseudoscalar and vector states are almost degenerate. Hence chiral symmetry does not apply to hadrons with charm or bottom quarks: The dynamics of the heavy quark component is governed by the heavy spin-flavor symmetry [106 , 107 , 108 ] while the light flavor content follows the laws of chiral symmetry.
To incorporate heavy flavor symmetry, two-component arrays P P P and Q ν Q_\nu Q ν are introduced that parameterize the pseudoscalar (mass M M M ) and vector mesons (mass M ∗ M^* M ∗ ) with a single heavy quark. The two components stand for the light quark, i.e., up or down content. This can be extended by a third component for the strange quark. These (almost degenerate) mesons are combined in the heavy meson multiplet
H = 1 2 ( 1 + γ μ V μ ) ( i γ 5 P ~ + γ α Q ~ α ) with P ~ = e i M V ⋅ x P and Q ~ α = e i M ∗ V ⋅ x Q α . (89) \mathcal{H} = \frac{1}{2}\left(1+\gamma_\mu V^\mu\right)\left(i\gamma_5\widetilde{P}+\gamma^\alpha\widetilde{Q}_\alpha\right)
\quad\text{with}\quad
\widetilde{P}=e^{iMV\cdot x}P
\quad\text{and}\quad
\widetilde{Q}_\alpha=e^{iM^*V\cdot x}Q_\alpha. \tag*{(89)} H = 2 1 ( 1 + γ μ V μ ) ( i γ 5 P + γ α Q α ) with P = e i M V ⋅ x P and Q α = e i M ∗ V ⋅ x Q α . (89)
The transformations P → P ~ P\to\widetilde{P} P → P and Q α → Q ~ α Q_\alpha\to\widetilde{Q}_\alpha Q α → Q α with four velocity V μ V_\mu V μ (V 2 = 1 V^2=1 V 2 = 1 ) define the reference frame for the heavy mesons such that the time derivative produces their kinetic energies. A model Lagrangian with the above postulated symmetries and as few as possible derivatives acting on the light pseudoscalar meson fields reads [109 ]
L ~ H = i M V μ Tr { H ( ∂ μ − i v μ ) H ‾ } − d Tr { H γ μ γ 5 p μ H ‾ } + … with p μ = i 2 ( ξ ∂ μ ξ † − ξ † ∂ μ ξ ) and v μ = i 2 ( ξ ∂ μ ξ † + ξ † ∂ μ ξ ) . (90) \widetilde{\mathcal{L}}_{\mathrm{H}}=iMV^\mu\operatorname{Tr}\left\{\mathcal{H}\left(\partial_\mu-iv_\mu\right)\overline{\mathcal{H}}\right\}-d\operatorname{Tr}\left\{\mathcal{H}\gamma_\mu\gamma_5p^\mu\overline{\mathcal{H}}\right\}+\ldots
\quad\text{with}\quad
p_\mu=\frac{i}{2}\left(\xi\partial_\mu\xi^\dagger-\xi^\dagger\partial_\mu\xi\right)
\quad\text{and}\quad
v_\mu=\frac{i}{2}\left(\xi\partial_\mu\xi^\dagger+\xi^\dagger\partial_\mu\xi\right) . \tag*{(90)} L H = i M V μ Tr { H ( ∂ μ − i v μ ) H } − d Tr { H γ μ γ 5 p μ H } + … with p μ = 2 i ( ξ ∂ μ ξ † − ξ † ∂ μ ξ ) and v μ = 2 i ( ξ ∂ μ ξ † + ξ † ∂ μ ξ ) . (90)
where the ellipsis indicate subleading pieces in 1 / M 1/M 1/ M and eventual interactions with light vector mesons, cf. Section Vector mesons . The single new parameter can be extracted from the semi-leptonic D → K D\to K D → K transition: d ≈ 0.53 d\approx0.53 d ≈ 0.53 [109 , 110 ] .15 Similar to the bound state approach in Section Extension to SU3 , substituting the hedgehog into p μ p_\mu p μ and v μ v_\mu v μ generates an attractive potential for the heavy mesons leading to a bound state whose extension is O ( 1 M ) O\left(\frac{1}{\sqrt{M}}\right) O ( M 1 ) . In the limit M → ∞ M\to\infty M → ∞ it is peaked at r = 0 r=0 r = 0 and the binding energy is given by the soliton profile(s) at the center:
E B = 3 2 d F ′ ( 0 ) − 3 2 g M G ′ ′ ( 0 ) + α 2 ω ( 0 ) E_B=\frac{3}{2}dF'(0)-\frac{3\sqrt{2}}{gM}G''(0)+\frac{\alpha}{2}\omega(0) E B = 2 3 d F ′ ( 0 ) − g M 3 2 G ′′ ( 0 ) + 2 α ω ( 0 )
[109 ] . The coupling constant c c c can be estimated from the semi-leptonic decay D → K ∗ D\to K^* D → K ∗ [110 ] to be c ≈ 1.60 c\approx1.60 c ≈ 1.60 . Unfortunately there is no direct empirical information about the parameter α \alpha α . Assuming light vector meson dominance for the electromagnetic form factors of the heavy mesons suggests α ≈ 1 \alpha\approx1 α ≈ 1 .
A more realistic treatment [111 ] starts with a Lorentz-invariant Lagrangian and demands that it approaches L ~ H \widetilde{\mathcal{L}}_{\mathrm{H}} L H when M = M ∗ → ∞ M=M^*\to\infty M = M ∗ → ∞ :
L H = ( D μ P ) † D μ P − 1 2 ( Q μ ν ) † Q μ ν − M 2 P † P + M ∗ 2 Q μ † Q μ + 2 i M d [ P † Q μ − Q μ † P ] p μ − d 2 ϵ α β μ ν [ Q α β † p μ Q ν + Q μ † p ν Q α β ] + … (91) \mathcal{L}_{\mathrm{H}}=(D^\mu P)^\dagger D_\mu P-\frac{1}{2}(Q^{\mu\nu})^\dagger Q_{\mu\nu}-M^2P^\dagger P+M^{*2}Q_\mu^\dagger Q^\mu+2iMd\left[P^\dagger Q^\mu-Q_\mu^\dagger P\right]p_\mu-\frac{d}{2}\epsilon^{\alpha\beta\mu\nu}\left[Q^\dagger_{\alpha\beta}p_\mu Q_\nu+Q_\mu^\dagger p_\nu Q_{\alpha\beta}\right]+\ldots \tag*{(91)} L H = ( D μ P ) † D μ P − 2 1 ( Q μν ) † Q μν − M 2 P † P + M ∗ 2 Q μ † Q μ + 2 i M d [ P † Q μ − Q μ † P ] p μ − 2 d ϵ α β μν [ Q α β † p μ Q ν + Q μ † p ν Q α β ] + … (91)
with D μ = ∂ μ − i v μ D_\mu=\partial_\mu-iv_\mu D μ = ∂ μ − i v μ and Q μ ν = D μ Q ν − D ν Q μ Q_{\mu\nu}=D_\mu Q_\nu-D_\nu Q_\mu Q μν = D μ Q ν − D ν Q μ . Again, terms with light vector mesons are not written out. The condition to assume Eq. (89) dictated the relative coefficients in the square brackets. Both S S S - and P P P -wave (as defined by the angular momentum of the pseudoscalar component) bound states have been determined numerically. The heavy limiting binding energy, E B E_B E B , in which the lowest energy bound states in these channels are degenerate, is approached only very slowly, even for M = M ∗ = 50 G e V M=M^*=50\,\mathrm{GeV} M = M ∗ = 50 GeV deviations of 10% (P P P -wave) or 20% (S S S -wave) are observed. Certainly, this limit does not apply for the charm sector. It is therefore mandatory to use the full model, Eq. (90) for a realistic description for baryons with a heavy quark. For two light flavors the Skyrme model soliton predicts their masses on the light side when compared to experiment, while the vector meson model yields fair agreement when α = 0 , … , 0.3 \alpha=0,\ldots,0.3 α = 0 , … , 0.3 [112 ] . When extending to three light flavors with exact diagonalization of the rigid rotator as in Section Extension to SU3 , substituting the Skyrme model soliton into p μ p_\mu p μ and v μ v_\mu v μ slightly overestimates the flavor symmetry breaking effects within a heavy baryon multiplet in the P P P -wave but predicts too small mass differences in the S S S -wave channel [113 ] ; though only a limited number of experimental data are available for the latter.
Beyond unit baryon number
So far we have only considered solitons describing hadrons with baryon number one. According to (37) the most suggestive procedure to construct solitons with baryon B ≥ 2 B\geq2 B ≥ 2 is to take F ( 0 ) = π F(0)=\pi F ( 0 ) = π . For n = 2 n=2 n = 2 the resulting classical energy is about three times as large as that of the B = 1 B=1 B = 1 hedgehog and thus exceeds the bound derived from (38) considerably. A smaller energy solution should exist. Indeed, doubling the velocity in the azimuthal angle, i.e., taking (29) and replacing
τ ⋅ x ^ = ( cos θ e − i φ sin θ e i φ sin θ − cos θ ) ⟶ ( cos θ e − 2 i φ sin θ e 2 i φ sin θ − cos θ ) \tau\cdot\hat{x}=
\begin{pmatrix}
\cos\theta& e^{-i\varphi}\sin\theta\\
e^{i\varphi}\sin\theta& -\cos\theta
\end{pmatrix}
\longrightarrow
\begin{pmatrix}
\cos\theta& e^{-2i\varphi}\sin\theta\\
e^{2i\varphi}\sin\theta& -\cos\theta
\end{pmatrix} τ ⋅ x ^ = ( cos θ e i φ sin θ e − i φ sin θ − cos θ ) ⟶ ( cos θ e 2 i φ sin θ e − 2 i φ sin θ − cos θ )
with a radially symmetric chiral angle and F ( 0 ) = π F(0)=\pi F ( 0 ) = π reduces that factor 3 to about 2.14 [114 ] . Allowing furthermore the chiral angle to additionally depend on the polar angle θ \theta θ the numeric (lattice) simulation finally produces bound configurations not only for B = 2 B=2 B = 2 but also for B = 3 , 4 , 5 B=3,4,5 B = 3 , 4 , 5 [115 ] . For B = 2 B=2 B = 2 the contour lines of equal mass density are (approximately) circles in planes containing the z z z -axis and that are centered at z = 0 z=0 z = 0 and a finite distance ρ 0 \rho_0 ρ 0 away from that axis. These circles form tori which are frequently called donuts. More recently these numeric simulations have been pushed to even produce bound configurations with baryon number larger than one hundred [116 ] . Whether or not these solutions should be identified with existing nuclei is not clear: the above argument that the eventually large O ( N C 0 ) O(N_C^0) O ( N C 0 ) quantum corrections can be ignored does not hold anymore, since they might vary substantially with baryon number [117 ] .
Another way to consider two nucleon systems is the so-called product ansatz
U 2 ( x , t ) = A 1 ( t ) U H ( x 1 ) A 1 † ( t ) A 2 ( t ) U H ( x 2 ) A 2 † ( t ) , (92) U_2(x,t)=A_1(t)U_{\mathrm{H}}(x_1)A_1^\dagger(t)A_2(t)U_{\mathrm{H}}(x_2)A_2^\dagger(t), \tag*{(92)} U 2 ( x , t ) = A 1 ( t ) U H ( x 1 ) A 1 † ( t ) A 2 ( t ) U H ( x 2 ) A 2 † ( t ) , (92)
where U H ( x ) U_{\mathrm{H}}(x) U H ( x ) is from Eq. (29) and represents a single baryon. The spatial separation is parameterized by x i = x + R / 2 x_i=x+R/2 x i = x + R /2 and A i ( t ) ∈ S U ( N f ) A_i(t)\in SU(N_f) A i ( t ) ∈ S U ( N f ) are the respective spin-flavor collective coordinates. From this ansatz the nucleon-nucleon potential can be extracted [118 , 119 , 120 ] by exploring the energy difference E [ U 2 ] − 2 E [ U H ] E[U_2]-2E[U_{\mathrm{H}}] E [ U 2 ] − 2 E [ U H ] as a function of ∣ R ∣ \lvert R\rvert ∣ R ∣ . The model reproduces the long-range one pion exchange attraction as expected from the asymptotic behavior, Eq. (32) . To account for the intermediate-range attraction (which is essential for binding nucleons to nuclei) the Skyrme model must contain L 6 L_6 L 6 [121 ] 16 or be augmented by scalar fields [121 , 123 ] ; on the other hand vector meson fields generate (short range) repulsion [124 ] .
For three flavors another B = 2 B = 2 B = 2 configuration was found from the ansatz [125 ]
U H D B = 1 3 × 3 e i ψ + i Λ ⋅ x ^ e − i ψ / 2 sin χ + ( Λ ⋅ x ^ ) 2 [ e − i ψ / 2 cos χ − e i ψ ] (93) U_{\mathrm{HDB}} = \mathbf{1}_{3 \times3} e^{i\psi} + i\boldsymbol{\Lambda} \cdot\hat{x} e^{-i\psi/2} \sin\chi+ \left(\boldsymbol{\Lambda} \cdot\hat{x}\right)^2 \left[e^{-i\psi/2}\cos\chi- e^{i\psi}\right] \tag*{(93)} U HDB = 1 3 × 3 e i ψ + i Λ ⋅ x ^ e − i ψ /2 sin χ + ( Λ ⋅ x ^ ) 2 [ e − i ψ /2 cos χ − e i ψ ] (93)
where ψ \psi ψ and χ \chi χ are radial functions and Λ = ( λ 7 , − λ 5 , λ 2 ) \boldsymbol{\Lambda} = (\lambda_7,-\lambda_5,\lambda_2) Λ = ( λ 7 , − λ 5 , λ 2 ) . This configuration has strangeness S = − 2 S = -2 S = − 2 and is thus a model for the H-dibaryon, explored earlier in the bag model [126 ] . This baryon can only decay into two Λ \Lambda Λ -hyperons. Yet, for the physical value of f π f_\pi f π it has a binding energy as large as about 1 GeV. Further estimates of quantum corrections, however, suggest that this binding energy could be much smaller [127 ] ; though the approximation underlying this estimate has been questioned in Ref. [117 ] .
A conceptually different way to look at large baryon numbers is to compactify the target space R 3 \mathbb{R}^3 R 3 to a sphere S 3 ( L ) S^3(L) S 3 ( L ) , where L L L is the radius of the sphere [128 , 129 ] . A unit baryon number solution in the compactified space then has global baryon number density ∝ L − 3 \propto L^{-3} ∝ L − 3 . The sphere is embedded in Minkowski space by introducing a second polar angle 0 ≤ μ ≤ π 0 \le\mu\le\pi 0 ≤ μ ≤ π , such that x ∈ R 4 x \in\mathbb{R}^4 x ∈ R 4 is parameterized as x α = L e π ( cos μ , r ^ sin μ ) x^\alpha= \frac{L}{e_\pi}(\cos\mu,\hat{r}\sin\mu) x α = e π L ( cos μ , r ^ sin μ ) . The chiral angle F ( r ) F(r) F ( r ) in Eq. (29) is replaced by f ( μ ) f(\mu) f ( μ ) . The field equations for the Skyrme model (without pion mass term) have the identity map f I ( μ ) = μ f_I(\mu) = \mu f I ( μ ) = μ as a unit baryon number solution. Its energy, E [ f I ] = 3 π 2 f π e 2 ( L + 1 L ) E[f_I] = 3\frac{\pi^2 f_\pi}{e^2}\left(L+\frac{1}{L}\right) E [ f I ] = 3 e 2 π 2 f π ( L + L 1 ) saturates the bound, mentioned after Eq. (38) , at its minimum, L = 1 L = 1 L = 1 . Hence for this radius the identity map is the true solution; but as L → ∞ L \to\infty L → ∞ the hedgehog from Figure 3 must be the true solution. It turns out that f I f_I f I is stable for L < 2 L < \sqrt{2} L < 2 but unstable for L > 2 L > \sqrt{2} L > 2 . Since tr U = 2 cos [ f ( μ ) ] \operatorname{tr}U = 2\cos[f(\mu)] tr U = 2 cos [ f ( μ )] , the VEV ⟨ tr U ⟩ \langle\operatorname{tr}U\rangle ⟨ tr U ⟩ vanishes for the identity map, i.e., for small L L L and high densities. Numerical studies [129 ] show that ⟨ tr U ⟩ \langle\operatorname{tr}U\rangle ⟨ tr U ⟩ , which can be viewed as the quark condensate (cf. the discussion after Eq. (82) ), is a continuous but not a differentiable function of L L L , evidencing that the phase transition is second order. This suggests to relate the observed phase transition to the chiral phase transition. Yet, the predicted critical density of about 0.19 f m − 3 0.19\,\mathrm{fm}^{-3} 0.19 fm − 3 [129 ] is lower than the QCD-related one which is expected to occur at several times the nuclear density which by itself is already 0.15 f m − 3 0.15\,\mathrm{fm}^{-3} 0.15 fm − 3 .
Conclusion
In this article we have reviewed the basic concepts of chiral solitons. We have discussed their motivation from the chiral symmetry of strong interactions and large-N C N_C N C QCD. We explored the Skyrme model for baryons in some detail. Especially we employed that model to explain the main concepts of and calculational techniques for chiral solitons; both for two and three light, or almost light, flavors. We also had a glance at refinements of the model, which improve on the agreement with empirical data. Though these model predictions are not extremely precise, it is in total fascinating to observe that the single and maybe simple concept of chiral solitons is able to capture almost all baryon physics, from the very small (structure functions), via the obvious (spectrum and static properties) to the very large (nuclei and dense matter).
This article has focused on chiral soliton models in the context of particle and nuclear physics. Yet, there are applications outside this realm. For example, the two (space) dimensional Skyrmions, called baby [130 ] or magnetic [131 , 132 ] Skyrmions, are highly relevant in condensed matter physics where they describe spin textures with topological charges [133 ] .
Acknowledgments
The author has gained his insight into the topic of chiral soliton models for baryons by collaborations with many colleagues: R. Alkofer, L. Gamberg, G. Holzwarth, Ulf-G. Meißner, H. Reinhardt, J. Schechter, B. Schwesinger, and H. Walliser. Their input is gratefully acknowledged. The author is supported in part by the NRF (South Africa) under grant 150672.
See Also: Chiral Perturbation Theory, Large N C N_C N C Hadron Physics