Three self-similar solutions of Yang-Mills equations in high odd dimensions Piotr Bizoń, Irfan Glogić, Arthur Wasserman Abstract . We consider spherically symmetric Yang-Mills equations with gauge group S ​ O ​ ( d ) SO(d) in d + 1 d+1 dimensional Minkowski spacetime. For any given odd d ≥ 11 d\geq 11 , we establish existence and uniqueness (modulo reflection symmetry) of exactly N N smooth self-similar solutions, where N N is the number of zeros of an explicit polynomial P m ​ ( z ) P_{m}(z) of degree m = ( d − 5 ) / 2 m=(d-5)/2 in the interval 0 < z < 1 0-\frac{d-2}{4} , hence y\sqrt{1-y^{2}}|u^{\prime}|<\frac{\sqrt{d-2}}{2} . It follows that u^{\prime}(y) and u(y) have finite limits at y=1 . Lemma 1 . If d\geq 10 , then u(y) is monotone decreasing from u(0)=1 to u(1)>0 . Proof. Let h(y)=y^{3}u^{\prime}(y) . Differentiating Eq. ( 5 ), we obtain y^{2}(1-y^{2})h^{\prime\prime}+(d-7)yh^{\prime}-\left(3(d-2)u(y)^{2}+d-10\right)h=0. (11) Since h(y)\sim-2ay^{4} near y=0 , it is negative and decreasing for small y . Suppose, for contradiction, that y_{0}<1 is the first point where h^{\prime}(y_{0})=0 . Evaluating Eq. ( 11 ) at y=y_{0} we get y_{0}^{2}(1-y_{0}^{2})h^{\prime\prime}(y_{0})=\left(3(d-2)u(y_{0})^{2}+d-10\right)h(y_{0})<0, (12) which contradicts the assumption that such a point y_{0} exists. Thus, h(y)<0 , hence u^{\prime}(y)<0 on (0,1) . By continuity, u^{\prime}(1)<0 . Evaluating Eq. ( 5 ) at y=1 , we obtain (d-5)u^{\prime}(1)+(d-2)u(1)(1-u(1)^{2})=0\,. (13) Since u^{\prime}(1)<0 and u(1)^{2}<1 , this implies that u(1)>0 . ∎ It follows from Lemma 1 that a general solution u(a,y) admits the Fuchsian expansion ( 9 ) at y=1 with c>0 . This defines a \mathcal{C}^{m-1} map a\mapsto c(a) from (0,\infty) to (0,1] . Lemma 2 . If d\geq 10 , then the function c(a) is monotone decreasing from c(0)=1 to c(\infty)=~0 . Proof. Obviously c(0)=1 . To analyze the limit a\rightarrow\infty , we introduce the rescaled variable e^{\tau}=\sqrt{a}y . Substituting u(y)=U(\tau) into Eq. ( 5 ) and taking the limit a\rightarrow\infty , we obtain the autonomous equation U^{\prime\prime}+(d-4)U^{\prime}+f(U)=0\,. (14) By an elementary phase plane analysis we find that solutions starting at (U,U^{\prime})=(1,0) at \tau=-\infty tend monotonically to (0,0) for \tau\rightarrow\infty . This proves that c(\infty)=0 . Next, let v(y)=y^{2}\dfrac{\partial u(a,y)}{\partial a} . Differentiating Eq. ( 5 ) with respect to a , we obtain y^{2}(1-y^{2})v^{\prime\prime}+\left((d-7)y+2y^{3}\right)v^{\prime}-\left(3(d-2)u(y)^{2}+d-10\right)v=0. (15) Since v(y)\sim-y^{4} near y=0 , it is negative and decreasing for small y . By the same argument as in the proof of Lemma 1, we conclude that v(y)<0 on (0,1) . By continuity, v(1)<0 , which proves that c(a) is monotone decreasing. ∎ Remark 3 . For 5\leq d\leq 9 , the assertions of Lemmas 1 and 2 are false because the solutions u(a,y) can oscillate around zero, and the function c(a) is not monotone (see [ 3 ] for the proof of existence of infinitely many smooth oscillating solutions in d=5 ). Thus, the condition d\geq 10 in these lemmas is sharp. Lemmas 1 and 2, combined with the above analysis of the behavior of smooth solutions at y=1 for odd d , imply that for d=2m+5\geq 11 the number of smooth solutions on [0,1] is equal (modulo the reflection symmetry) to the number N of zeros of the polynomial P_{m}(c^{2}) in the interval 0T . In terms of x Eq. ( 5 ) takes the form (1-x^{2})u^{\prime\prime}(x)+(d-5)xu^{\prime}-(d-2)u(1-u^{2})=0. (18) We showed above that this equation has three regular solutions in the interval 1\leq x<\infty , corresponding to the interior of the past light cone. The two explicit solutions u_{\pm}(x) obviously extend smoothly to the entire spacetime. That the third solution u_{\ast}(x) can be extended to x\in(-1,1) can be easily shown using the functional \tilde{H}(x)=-H(1/x)=\frac{1}{2}(1-x^{2})u^{\prime}(x)^{2}+\frac{d-2}{4}(1-u(x)^{2})^{2}, (19) which satisfies \tilde{H}^{\prime}(x)=-(d-4)xu^{\prime 2} . For x\rightarrow-1^{+} , in analogy to ( 9 ), we have u_{\ast}(x)=\tilde{c}+\sum_{n=1}^{m}\tilde{c}_{n}(x+1)^{n}+\tilde{b}(x+1)^{m+1}\log(x+1)+\tilde{c}_{m+1}(x+1)^{m+1}+\dots, (20) with a generically nonzero coefficient \tilde{b} , hence u_{\ast}(x) is only \mathcal{C}^{m} at x=-1 . Outlook Each smooth self-similar solution of Eq. ( 3 ) provides an example of finite-time singularity formation (“blowup”) from smooth initial data. The extent to which a given self-similar solution participates in the general dynamics depends on its stability. The stability of the solution u_{+} for all d\geq 5 was proved by the second author [ 7 , 13 ] , building on earlier work by the first author [ 8 ] and by Donninger et al. [ 9 , 10 ] (see also recent extensions of this result beyond the past light cone [ 11 , 12 ] ). This represents an important step toward showing that the self-similar solution u_{+}(\frac{r}{T-t}) governs generic spherically-symmetric blowup, as conjectured by the first author and supported by numerical computations [ 3 , 5 , 6 ] . The stability properties of the solutions u_{-} and u_{\ast} have not yet been established. Our spectral stability analysis indicates that the solution u_{-} has two unstable modes for 11\leq d\leq 17 and one unstable mode for d\geq 19 , while the solution u_{\ast} has one unstable mode for 11\leq d\leq 17 and two unstable modes for d\geq 19 . The solution with a single unstable mode (either u_{\ast} or u_{-} , depending on the dimension) is expected to be critical, in the sense that its codimension-one stable manifold separates regular and singular solutions (see [ 5 ] for numerical computations of critical dynamics in d=5 ). Finally, we point out that for d\geq 10 one can construct non-self-similar blowup solutions, in analogy with the construction of such solutions for equivariant wave maps in d\geq 7 [ 14 ] . For d\geq 11 these so-called type II blowup solutions have the form Q\left(r/\lambda(t)\right) , where Q(r) is a static solution of Eq. ( 3 ) and \lambda(t)\sim(T-t)^{1+\gamma} . The anomalous exponent \gamma>0 can be determined by the method of matched asymptotics. The coexistence of type I and type II blowup scenarios for d\geq 10 may lead to new blowup dynamics that are absent in lower dimensions 1 1 1 Let us recall that in the energy critical case d=4 the blowup must be of type II [ 15 , 16 ] , while for 5\leq d\leq 9 the blowup must be self-similar. . Acknowledgement. The research of P.B. was supported in part by the Polish National Science Center grant no. 2017/26/A/ST2/00530. The research of I.G. was funded in whole or in part by the Austrian Science Fund (FWF) 10.55776/PAT5825523. Appendix To simplify the notation, we introduce the variable z=c^{2} . In order to count the number of zeros of the polynomials P_{m}(z) for m\geq 3 , let us first factor out the two explicit zeros corresponding to the solutions u_{\pm} , z_{\pm}=\left(1-\frac{\alpha_{\pm}}{\beta_{\pm}+1}\right)^{2}. (21) Accordingly, we write P_{m}(z)=(z-z_{-})(z-z_{+})S_{m}(z), where S_{m}(z)=s_{m-2}z^{m-2}+s_{m-3}z^{m-3}+\cdots+s_{0}. (22) For any finite m , the coefficients s_{0},\dots,s_{m-2} can be computed explicitly in an algorithmic manner. By convention, we assume that the leading coefficient s_{m-2} is positive. The key observation (verified for increasingly large m , as if there was no tomorrow) is that the leading coefficient s_{m-2} is sufficiently large and all other coefficients s_{m-3},\dots,s_{0} are negative. Consequently, S_{m}(0)<0 and S_{m}(1)>0 . Since S_{m}(z) can change sign only once, it follows that is has a unique zero z_{\ast}\in(0,1) . We find that z_{-}