--- title: Type II Blowup in Nonlinear Equations url: https://www.emergentmind.com/topics/type-ii-blowup type: topic --- # Type II Blowup in Nonlinear Equations Searching arXiv for the cited papers to ground the article and confirm metadata. arxiv.search query: 2002.05765 Type II blowup is a classification of finite-time singularity formation in nonlinear evolution equations that identifies blowup not governed by the canonical self-similar, ODE, or scale-invariant regime. In the semilinear heat equation \(u_t=\Delta u+|u|^{p-1}u\), the reference rate comes from the ODE \(v'=v^p\), namely \(v(t)\sim C(T-t)^{-1/(p-1)}\); blowup outside this regime is called Type II. In the three-dimensional energy-critical heat equation \(u_t=\Delta u+u^5\), the rigorous construction of solutions with \(\|u(\cdot,t)\|_{L^\infty}\sim (T-t)^{-k}\), \(k=1,2,\dots\), provided the first instance of Type II finite-time blowup in that model. In other equations the operational definition changes—most notably, in Navier–Stokes it is phrased through divergence of scale-invariant Caffarelli–Kohn–Nirenberg quantities—but the unifying feature is a singularity that cannot be reduced to the natural scaling law of the underlying flow [2002.05765, 1201.2907, 2606.29468]. ## 1. Definitions and scaling frameworks For the Fujita-type heat equation \[ u_t=\Delta u+|u|^{p-1}u, \] finite-time blowup at \(T<\infty\) means \[ \lim_{t\to T^-}\|u(\cdot,t)\|_{L^\infty(\mathbb R^n)}=+\infty. \] The ODE model \(v'=v^p\) yields \[ v(t)\sim C(T-t)^{-\frac{1}{p-1}}, \] and this sets the natural Type I benchmark. A blowup is Type I if \[ \limsup_{t\to T^-}(T-t)^{\frac{1}{p-1}}\|u(\cdot,t)\|_{L^\infty}<+\infty, \] and Type II if this quantity is infinite. For the three-dimensional critical exponent \(p=5\), the ODE rate is \((T-t)^{-1/4}\), so any regime with \(\|u\|_\infty\sim (T-t)^{-k}\), \(k\ge 1\), is decisively Type II [2002.05765]. In the four-dimensional energy-critical heat equation, the same distinction is often expressed through the blowup scale \(\lambda(t)\). Type I corresponds to the self-similar scale \(\lambda(t)\sim \sqrt{T-t}\), whereas Type II means \(\lambda(t)\ll \sqrt{T-t}\). The amplitude \(1/\lambda(t)\) may then be larger than the Type I amplitude even though the classification is made at the level of the scaling law rather than a single norm. This makes clear that “Type II” is not a universal inequality in one fixed norm, but a deviation from the canonical scaling mechanism of the equation under study [1201.2907]. For Navier–Stokes, the terminology is different. At a singular point \(z=0\), one considers scale-invariant quantities such as \[ A(v,r)=\sup_{-r^20\), there exists smooth initial data such that the corresponding solution blows up at time \(T\) and satisfies \[ \|u(\cdot,t)\|_{L^\infty(\mathbb R^3)}\sim (T-t)^{-k},\qquad k=1,2,\dots. \] This gives a rigorous realization of the formal asymptotics predicted by Filippas–Herrero–Velázquez in dimension \(3\), and it is the first rigorous construction of Type II finite-time blowup for the three-dimensional energy-critical heat equation [2002.05765]. The construction is radial and produces blowup at a single point, the origin. The solution is asymptotically a matched composite of a concentrating rescaled ground state in the inner region and a global outer component governed by self-similar heat dynamics. The relevant bubble scale \(p(t)\) satisfies \[ p(t)\sim p_0(t):=3^{1/2}2A(T-t)^{2k}, \] so the bubble radius shrinks like \((T-t)^{2k}\). Since the amplitude near the origin is of order \(p(t)^{-1/2}(T-t)^{-1/2}\), the \(L^\infty\)-norm indeed behaves like \((T-t)^{-k}\) [2002.05765]. ## 3. Blowup profile, matching, and the inner–outer mechanism The inner profile is built from the stationary solution of \[ -\Delta W=W^5. \] Near the blowup point, the solution is modeled by a concentrating bubble \(w(x/p(t))\), together with a correction driven by the scaling mode. The correction \(J\) solves \[ \Delta J+5w^4J+Z_0=0, \] where \[ Z_0(y)= -\bigl(y\cdot \nabla w+\tfrac12 w\bigr) \] is the generator of scaling. The function \(Z_0\) decays like \(r^{-1}\) and is not in \(L^2\), and this slow decay is a structural source of the nonlocal modulation law [2002.05765]. The outer profile is organized in self-similar variables \[ z=\frac{x}{\sqrt{T-t}},\qquad \tau=-\log(T-t),\qquad \Phi(z,\tau)=(T-t)^{1/2}u(x,t), \] for which the linearized outer equation is \[ \partial_\tau\Phi=\Delta_z\Phi-\frac12 z\cdot\nabla_z\Phi-\frac12\Phi. \] Stationary or exponentially weighted modes of this operator are expressed through Hermite polynomials. Choosing \[ \gamma=\frac12-k,\qquad k\in\mathbb Z_+, \] one obtains polynomially growing outer modes of the form \[ m(z)=A\,C_k\,H_{2k}(|z|), \] hence, in original variables, \[ u_{\mathrm{out}}(x,t)=A(T-t)^{k-\frac12}C_k\,H_{2k}\!\left(\frac{|x|}{\sqrt{T-t}}\right). \] These modes dominate on the self-similar scale \(|x|\sim \sqrt{T-t}\) [2002.05765]. The decisive step is matching in the intermediate region \[ p(t)\ll |x|\ll \sqrt{T-t}. \] There the inner expansion has leading behavior proportional to \(p(t)/|x|\), while the outer Hermite mode contributes a term proportional to \((T-t)^k/|x|\). Matching these two asymptotics forces \[ p(t)\sim \text{const}\cdot (T-t)^{2k}. \] The Type II law is therefore not imposed externally; it is selected by the compatibility of the inner stationary bubble and the outer linear heat mode. This mechanism is the canonical bubbling picture in the three-dimensional critical heat flow: a rescaled ground state, a global Hermite mode, and a modulation law extracted from matching [2002.05765]. ## 4. Gluing, spectral theory, and the reduced nonlocal dynamics The analytic construction proceeds through an inner–outer gluing scheme. In self-similar variables one studies \[ \partial_\tau \Phi=\Delta_z\Phi-\frac12 z\cdot\nabla_z\Phi-\frac12\Phi+\Phi^5, \] while in the inner region one introduces \[ y=\frac{x}{p(t)},\qquad p(t)=p_0(t)(1+A(t))^2. \] The perturbation near the bubble is written as a correction \(\phi(y,t)\), whereas the global remainder is an outer term \(\psi(x,t)\). The total perturbation is decomposed as \[ u=U_1+\Phi_1+\Phi_2,\qquad \Phi_2=\psi(x,t)+\eta\!\left(\frac{x}{Rp(t)}\right)\phi\!\left(\frac{x}{p(t)},t\right), \] with \(\eta\) a cutoff [2002.05765]. The linearized inner operator is \[ L\phi:=\Delta\phi+5w^4\phi. \] Its spectral structure has one negative radial eigenvalue and the slow-decaying zero mode \(Z_0\). The inner linear theory in three dimensions therefore requires orthogonality to \(Z_0\), \[ \int_{B_{2R}} h(y,t)Z_0(y)\,dy=0\qquad\forall t, \] in order to invert the operator on weighted spaces. The outer problem is solved by weighted parabolic estimates for \[ \partial_t\psi=\Delta\psi+G(\phi,\psi,a), \] using Duhamel representation and heat kernel bounds [2002.05765]. The central modulation law comes from the solvability condition \[ \int_{B_{2R}} H(\phi,\psi,a)Z_0(y)\,dy=0. \] After asymptotic expansion, this produces a reduced nonlocal equation for the scale modulation \(a(t)\), \[ \int_0^t \frac{a(s)}{(t-s)^{1/2}}\,ds = \sum_{j=1}^k c_j B^{(j)}(0,t)+h(t). \] This equation is essentially a Caputo fractional derivative of order \(1/2\). Lemma 5.1 in the construction shows that, for suitable choices of constants \(c_j\) and auxiliary heat profiles \(B^{(j)}\), one can solve it with \[ a(t)\sim (T-t)^{k-1}A(t),\qquad A(t)\to 0\quad\text{as }t\to T. \] Through the relation between \(a(t)\) and \(p(t)\), this yields \[ p(t)\sim (T-t)^{2k},\qquad \|u(\cdot,t)\|_{L^\infty}\sim (T-t)^{-k}. \] Existence is then closed by a compact self-mapping argument in a Banach space for \((\phi,\psi,A,c)\), with Schauder fixed point theorem furnishing the desired solution [2002.05765]. ## 5. Comparative manifestations across nonlinear PDE The heat equation provides the most classical setting for Type II blowup, but the phenomenon is broader and structurally diverse. In the four-dimensional energy-critical semilinear heat equation, a radial Type II blowup concentrates the Talenti–Aubin soliton \[ Q(r)=\frac{1}{1+r^2/8} \] with scale \[ \lambda(t)=c(u_0)\frac{T-t}{|\log(T-t)|^2}(1+o(1)), \] and convergence \[ u(t,x)=\frac{1}{\lambda(t)}Q\!\left(\frac{x}{\lambda(t)}\right)+u_*(x)+o_{H^1}(1). \] This is Type II because \(\lambda(t)\ll \sqrt{T-t}\), even though the asymptotic description is again a single universal bubble plus radiation [1201.2907]. In the five-dimensional critical heat equation, type II solutions can blow up at finitely many prescribed points, \[ u(x,t)\approx \sum_{j=1}^k U_{\mu_j(t),\xi_j(t)}(x)+Z^*(x,t), \qquad \mu_j(t)=B_j (T-t)^2(1+o(1)), \] so that \[ \|u(\cdot,t)\|_{L^\infty}\sim (T-t)^{-3}. \] This extends the bubbling mechanism from one point to multiple points, with each core modeled by an Aubin–Talenti bubble [1808.10637]. Related constructions appear in other critical or supercritical flows. For the 1-corotational energy supercritical harmonic heat flow, one has \[ u(r,t)\sim Q\!\left(\frac{r}{\lambda(t)}\right),\qquad \lambda(t)\sim c_u(T-t)^{\ell/\gamma}, \] with quantized rates and \((\ell-1)\)-codimension stability, and the case \(\ell=1\) is stable. In the energy supercritical nonlinear Schrödinger equation, Type II blowup concentrates a solitary wave while all Sobolev norms below the scaling index remain bounded, and the scale again shrinks at quantized rates. In the three-dimensional axisymmetric Keller–Segel system, Type II blowup occurs along a ring and is locally modeled by the two-dimensional stationary profile \[ U(x)=\frac{8}{(1+|x|^2)^2}, \qquad \lambda(t)=\sqrt{T-t}\,\exp\!\left(-\sqrt{\frac{|\log(T-t)|}{2}+O(1)}\right). \] These examples indicate that Type II blowup is often a bubbling or soliton-concentration phenomenon rather than a self-similar one [1611.08877, 1407.1415, 2502.19775]. A different use of the term occurs in Navier–Stokes, where potential Type II blowup is analyzed through Euler-scaled limits. There the strategy is to extract a nontrivial ancient Euler solution from a hypothetical singularity and then exclude it by Liouville-type arguments. This does not construct a blowup, but it places Type II squarely at the interface between critical rescaling and inviscid limiting dynamics [2606.29468]. ## 6. Stability, codimension, and open directions Type II blowup is frequently misunderstood as a generic strengthening of Type I. The available constructions indicate the opposite. In the three-dimensional energy-critical heat equation, the blowup solutions are obtained by delicate parameter selection and orthogonality conditions; the paper does not prove stability and makes no uniqueness claim. The main radial theorem produces one-point blowup at the origin, and only a sketch of nonradial multi-point blowup is given, then only for the basic rate \(k=1\) [2002.05765]. By contrast, the four-dimensional critical heat flow provides a codimension-one picture: the Type II regime lies on a codimension-one manifold of radial initial data, reflecting a single unstable direction in the linearized dynamics. In the harmonic heat flow analogue, the \(\ell=1\) regime is stable while higher \(\ell\) are higher-codimension. These comparisons suggest that the integer \(k\) or \(\ell\) indexing a Type II rate is often also an instability index, with faster or more elaborate blowup laws requiring more tuning of the initial data [1201.2907, 1611.08877]. Several open problems recur across the literature. Stability under perturbations remains unresolved in the three-dimensional critical heat equation. Classification is largely open: it is not known whether every Type II blowup in that equation must arise from bubbling of the ground state with a discrete rate, or whether more exotic patterns exist. Nonradial and multi-bubble constructions beyond the first rate, and the interaction of several bubbles or several blowup points, remain technically difficult. More generally, the recurring presence of slow-decaying kernel modes, fractional modulation laws, and geometry-dependent concentration suggests that Type II blowup is best understood not as a single phenomenon but as a family of singular regimes in which stationary or soliton-like structures dominate the asymptotics and force the evolution away from naive self-similarity [2002.05765, 1201.2907]. Source: https://www.emergentmind.com/topics/type-ii-blowup