Derrick Scaling, Virial Tests, and Nontopological Stability

Derrick scaling tests whether a proposed static localized field can be stationary under a uniform change of size. If every positive contribution to the energy changes with the same sign, no stationary finite-size lump exists under the theorem’s hypotheses. An evasion must supply a term with the opposite scaling, impose a conserved-charge constraint, change the field content or geometry, or leave the static ansatz through genuine time dependence. Naming the stabilizer is part of the result.

Required background. Variational field equations and conserved currents supplies the stationary-action argument, and interactions, potentials, and stability supplies the bounded-energy assumptions. Helpful background. EFT power counting is needed when higher-derivative stabilization is used near a cutoff.

The scaling identity

Consider $n$ real scalar fields on flat $\mathbb R^D$ with static energy

$$E[\phi]=E_2+E_0 =\int\mathrm d^D x \left[ \frac12 G_{ab}(\phi)\, \partial_i\phi^a\partial_i\phi^b+V(\phi) \right],$$

where $G_{ab}$ is positive definite, $V\geq0$ after subtracting the vacuum energy, and the fields approach a vacuum fast enough for all integrals and integrations by parts to exist. Define the size variation

$$\phi_\lambda(\mathbf x)=\phi(\lambda\mathbf x), \qquad \lambda>0 .$$

Changing variables to $\mathbf y=\lambda\mathbf x$ gives

$$E(\lambda) =\lambda^{2-D}E_2+\lambda^{-D}E_0 .$$

Any regular static solution is stationary under this admissible variation, so

$$\left.\frac{\mathrm dE}{\mathrm d\lambda}\right|_{\lambda=1} =(2-D)E_2-DE_0=0 .$$

This is the virial identity. For $D\geq3$, both coefficients are negative and a nontrivial solution with $E_2,E_0\geq0$ is impossible. In $D=2$, stationarity requires $E_0=0$ and leaves the two-derivative energy scale invariant; size is not fixed. In $D=1$, the relation $E_2=E_0$ permits a kink. This is the core of Derrick 1964, pp. 1252–1254; Coleman 1985, § 2, pp. 194–195 gives the field-theory interpretation.

More generally, if a positive energy term contains a total of $k$ spatial derivatives and its field amplitudes are not rescaled, then

$$E_k[\phi_\lambda]=\lambda^{k-D}E_k[\phi].$$

For $E=\sum_k E_k$, the scale residual is

$$\mathcal R_{\mathrm D} \equiv \left.\frac{\mathrm dE(\lambda)}{\mathrm d\lambda}\right|_1 =\sum_k(k-D)E_k .$$

A numerical profile claimed to be static should satisfy $\mathcal R_{\mathrm D}=0$ within a tolerance controlled separately from the field-equation residual. A small equation residual on a finite grid can coexist with a large virial error if the box, tails, or boundary data are wrong.

Hypotheses that must travel with the obstruction

The conclusion is not “solitons do not exist above one spatial dimension.” It concerns a particular variational problem:

• the configuration is time independent and localized on unbounded flat space;
• the scale variation stays within the admissible field space and preserves the declared boundary sector;
• the energy consists of the stated positive terms and has no explicit position-dependent scale;
• the fields are regular enough that $E(\lambda)$ is differentiable at $\lambda=1$;
• no conserved quantity is being held fixed by an additional Lagrange multiplier;
• boundary, curvature, background fields, gauge constraints, and higher derivatives have not been omitted.

If any item fails, the scaling calculation must be redone rather than cited away. The theorem is a diagnostic: it says exactly what kind of new contribution is needed. Manton and Sutcliffe 2004, § 4.2, pp. 82–87 discuss the general scaling method and its model dependence.

A controlled evasion: the Skyrme balance

Let $U(\mathbf x)\in SU(2)$ in three spatial dimensions and define $L_i=U^\dagger\partial_iU$. A standard static Skyrme energy has two- and four-derivative pieces,

$$E_2 =\frac{f_\pi^2}{16} \int\mathrm d^3x\, \big[-\operatorname{tr}(L_iL_i)\big],$$ $$E_4 =\frac{1}{32e^2} \int\mathrm d^3x\, \big[-\operatorname{tr}([L_i,L_j][L_i,L_j])\big],$$

with positive integrands for anti-Hermitian $L_i$. A nonnegative potential contribution $E_0$ may also be present. Under $U_\lambda(\mathbf x)=U(\lambda\mathbf x)$,

$$E(\lambda)=\lambda^{-1}E_2+\lambda E_4+\lambda^{-3}E_0 ,$$

so a stationary solution must satisfy

$$-E_2+E_4-3E_0=0 .$$

Without $E_4$, shrinking lowers the two-derivative energy and no finite size is selected. The four-derivative term grows under shrinking and can balance $E_2$ and $E_0$. This is a real evasion because it changes the energy functional; topology alone did not do the work. The original higher-derivative construction is due to Skyrme 1961, pp. 127–138, while the modern scaling analysis is summarized in Manton and Sutcliffe 2004, §§ 4.2 and 9.1, pp. 82–87 and 349–356.

If the four-derivative operator belongs to an EFT with cutoff $\Lambda$, the solution’s inverse size must remain parametrically below $\Lambda$, and omitted operators must be smaller on the profile. A formal virial balance at $R^{-1}\sim\Lambda$ is not a controlled EFT prediction.

Other mechanisms and their tests

Gauge fields. Gauge potentials carry their own scaling required by covariance, and magnetic-field energy can balance Higgs-gradient and potential terms. The smooth monopole is the canonical example. One must scale the full gauge–Higgs ansatz; applying the scalar formula only to the Higgs profile gives the wrong conclusion.

Fixed Noether charge. A Q-ball has $\Phi(t,\mathbf x)=e^{i\omega t}f(\mathbf x)$. It is stationary only in energy density, not as a field. The relevant variational problem minimizes $E$ at fixed $Q$, or equivalently extremizes $E-\omega Q$. The time-dependent phase contributes a term absent from the static scalar theorem.

Periodic or rotating motion. Oscillons and rotating solitons can evade a static obstruction dynamically. Their claim is then a lifetime, radiation rate, or orbital stability statement, not a static minimum.

Boundaries and curvature. A finite box, compact space, impurity, or curved metric introduces a length and changes the scaling variation. A profile stabilized by its container is not automatically a soliton of the infinite-volume theory.

Higher derivatives or noncanonical kinetics. These alter the exponents and may stabilize a size, but positivity, hyperbolicity, additional modes, and EFT control must be checked independently.

The shared stability taxonomy shows why passing the virial test is necessary but never sufficient: existence, the Hessian spectrum, nonlinear evolution, and quantum corrections remain separate. The family-by-family scaling mechanisms, fluctuation claims, and failure boundaries are collected in the soliton boundary and stability comparison.

From virial stationarity to stability

The second scale derivative gives information only along the dilation direction. For the two-plus-four-derivative example in $D=3$,

$$\left.\frac{\mathrm d^2E}{\mathrm d\lambda^2}\right|_1 =2E_2+12E_0>0 ,$$

after using the virial identity. The solution is stable against infinitesimal uniform rescaling. It may still have a negative fluctuation with a different shape, so the full Hessian must be studied:

$$\mathcal H_{ab}(\mathbf x,\mathbf y) =\left. \frac{\delta^2E}{\delta\phi^a(\mathbf x)\delta\phi^b(\mathbf y)} \right|_{\phi_{\mathrm{sol}}}.$$

Zero eigenvalues generated by exact symmetries are collective coordinates, not instabilities. A negative eigenvalue is an unstable direction. Positive spectrum apart from normalizable symmetry zero modes establishes linear stability, not a general theorem of nonlinear or quantum stability.

A practical scaling test

For an analytic ansatz or numerical profile:

1. list every contribution to the conserved energy, including boundary and constraint terms;
2. declare how every field scales so that gauge covariance and boundary data are preserved;
3. compute each exponent and the residual $\mathcal R_{\mathrm D}$;
4. check the field equations and $\mathcal R_{\mathrm D}$ independently;
5. identify the pair of terms that prevents shrinking and spreading;
6. verify that the stabilizing term is inside its domain of validity;
7. only then compute the full fluctuation spectrum.

Stop if the scale variation changes charge, leaves the admissible field space, or probes an EFT cutoff. In those cases the naive residual is not a valid test.

Common pitfalls

Treating a vanishing virial residual as a solution. It is one integrated consequence of the field equations. Many nonsolutions can satisfy it accidentally.

Applying the scalar theorem to a gauge–Higgs system. Gauge covariance fixes how the gauge potential scales. Omitting its energy or rescaling it inconsistently invalidates the identity.

Using an uncontrolled higher-derivative term as a stabilizer. A new term can balance the scaling while every still-higher operator is equally important. The solution is predictive only if its gradients remain below the cutoff and the truncation is demonstrably ordered.

Exercises

1. For $E(\lambda)=\lambda^{2-D}E_2+\lambda^{-D}E_0$, derive the stationarity condition and classify $D=1,2,3$ when both contributions are nonnegative.

Solution

Differentiation at $\lambda=1$ gives $(2-D)E_2-DE_0=0$. For $D=1$, this is $E_2=E_0$ and permits a balance. For $D=2$, it requires $E_0=0$, leaving scale invariance of $E_2$. For $D=3$, it reads $-E_2-3E_0=0$, so only the vacuum with both terms zero is possible under the hypotheses.

2. In $D=3$, suppose $E=E_2+E_4$ with positive two- and four-derivative terms. Find the stationary relation and show that the dilation mode is locally stable.

Solution

The scaled energy is $E(\lambda)=\lambda^{-1}E_2+\lambda E_4$. Stationarity gives $E_4=E_2$. The second derivative is $2E_2>0$. This proves positivity only in the uniform-scale direction; the remaining fluctuation spectrum is not determined.

Continue

Lumps, Textures, and Skyrmions applies the test to scale-invariant and higher-derivative models. Q-Balls, Oscillons, and Sphalerons compares constrained, dynamical, and unstable-saddle evasions.

References

• Coleman, Sidney. Aspects of Symmetry: Selected Erice Lectures. Cambridge University Press, 1985, ch. 6, pp. 185–264. DOI.
• Derrick, G. H. “Comments on Nonlinear Wave Equations as Models for Elementary Particles.” Journal of Mathematical Physics 5 (1964): 1252–1254. DOI.
• Manton, Nicholas, and Paul Sutcliffe. Topological Solitons. Cambridge Monographs on Mathematical Physics. Cambridge University Press, 2004, §§ 4.2 and 9.1. DOI.
• Skyrme, T. H. R. “A Non-Linear Field Theory.” Proceedings of the Royal Society of London A 260 (1961): 127–138. DOI.