Charged scalar fields on Reissner–Nordström spacetimes II: late-time tails and instabilities

Let r^{-1}\psi be a solution to (?) with \mathfrak{q}\neq 0 with respect to the conformally smooth electromagnetic gauge A=-\frac{Q}{r}d\tau, with \tau a time function whose level sets are horizon-intersecting and asymptotically hyperboloidal spacelike hypersurfaces and with \partial_{\tau} a Killing vector field normalized at infinity that is timelike away from the event horizon. Assume that \psi arises from the evolution of smooth, compactly supported initial data. Then, for 1-\frac{|Q|}{M}\ll 1 or |\mathfrak{q}Q|\ll 1:

where \Psi^{+}_{\ell m}(\tau,r;\mathfrak{q},\mathfrak{H}_{\ell m}[\psi])=e^{-i\mathfrak{q}Qr_+^{-1}\tau}\Psi_{\ell m}^{\infty}\left(\tau,r_++\frac{r_+^2}{r-r_+};-\mathfrak{q},\mathfrak{H}_{\ell m}[\psi]\right), with r_+ the event horizon area radius, and where \Psi_{\ell m}^{\infty} are globally-defined functions satisfying the following late-time asymptotics with r_0\geq r_+:

Here,

• \beta_{\ell}:=\sqrt{(2\ell+1)^2-4\mathfrak{q}^2Q^2}\in [0,1)\cup i(0,\infty),

• \sigma=\,\textnormal{sign}\,(\mathfrak{q}Q), \zeta=\zeta(\mathfrak{q}Q,\beta_{\ell}) is a complex constant (defined in Theorem 4.1) and \Gamma(\cdot) is the gamma function,

• \mathfrak{I}_{\ell m}[\psi] and \mathfrak{h}_{\ell m}[\psi] are linear functionals of (derivatives of) initial data at \tau=0, so that each \mathfrak{I}_{\ell m}[\psi] vanishes only for a codimension-1 set of initial data, \mathfrak{H}_{\ell m}[\psi]\equiv 0 when |Q|<M and in the case |Q|=M, each \mathfrak{H}_{\ell m}[\psi] also vanishes only for a codimension-1 set of initial data,

• \mathfrak{w}_{\ell}Y_{\ell m} are time-independent solutions to (?) that are smooth at r=r_+ and satisfy \mathfrak{w}_{\ell}(r_+;\mathfrak{q}Q)=1.

In the case that 1-\frac{|Q|}{M}\gtrsim 1 and |\mathfrak{q}Q|\gtrsim 1, the above results remain valid under the condition of the validity of mode stability on the real axis.

Under the assumptions of Theorem 1.1, the following asymptotic instability and non-decay results hold:

1. For all k\in \mathbb{N}_0, there exist constants \mathfrak{c}_k>0, such that for all \ell\in \mathbb{N}_0 with \ell(\ell+1)<\mathfrak{q}^2Q^2 and m\in \mathbb{Z} with |m|\leq \ell:

   \begin{aligned} \limsup_{\tau\to \infty}(1+\tau)^{-k-\frac{1}{2}+\frac{\textnormal{Re}\, \beta_{\ell}}{2}}|D_X^{k+1}\psi_{\ell m}|(\tau,r_+)\geq &\:|\mathfrak{c}_k| |\mathfrak{H}_{\ell m}[\psi]| \quad &(|Q|=M,\:\beta_{\ell}\neq 0),\\ \limsup_{\tau\to \infty}\log(1+\tau)(1+\tau)^{-k-\frac{1}{2}}|D_X^{k+1}\psi_{\ell m}|(\tau,r_+)\geq &\:|\mathfrak{c}_k| |\mathfrak{H}_{\ell m}[\psi]|\quad &(|Q|=M,\:\beta_{\ell}=0),\\ \limsup_{\tau\to \infty}(1+\tau)^{-k-\frac{1}{2}+\frac{\textnormal{Re}\, \beta_{\ell}}{2}}|(r^2D_X)^{k+1}\psi_{\ell m}|(\tau,\infty)\geq &\:|\mathfrak{c}_k| |\mathfrak{I}_{\ell m}[\psi]| \quad &(\beta_{\ell}\neq 0),\\ \limsup_{\tau\to \infty}\log(1+\tau)(1+\tau)^{-k-\frac{1}{2}}|(r^2D_X)^{k+1}\psi_{\ell m}|(\tau,\infty)\geq &\:|\mathfrak{c}_k| |\mathfrak{I}_{\ell m}[\psi]|\quad &(\beta_{\ell}=0), \end{aligned}

   with X a radial vector field tangential to \Sigma_{\tau}, the level sets of \tau.

2. When |\mathfrak{q}Q|>\frac{1}{2} (\beta_0\in i(0,\infty)), energy concentration occurs near \mathcal{H}^+ in the case |Q|=M and, with additional r-weights, also near \mathcal{I}^+ when |Q|\leq M, in the following sense: there exists a constant \mathfrak{c}_{\rm int}>0, such that for all r_+\leq r_H<r_I<\infty:

   \begin{aligned} \limsup_{\tau\to \infty}\int_{\Sigma_{\tau}\cap\{r\leq r_H\}} \mathbb{T}^{SF}[g_{M,Q},r^{-1}\psi](\mathbf{n}_{\tau},\mathbf{n}_{\tau})\,d\mu_{\Sigma_{\tau}}\geq &\: \mathfrak{c}_{\rm int}\sum_{\substack{\ell\in \mathbb{N}_0\\ \ell+\frac{1}{2}<|\mathfrak{q}Q|}}\sum_{\substack{m\in \mathbb{Z}\\ |m|\leq \ell}}|\mathfrak{H}_{\ell m}[\psi]|^2,\\ \lim_{\tau\to \infty}\int_{\Sigma_{\tau}\cap\{r> r_H\}} \mathbb{T}^{SF}[g_{M,Q},r^{-1}\psi](\mathbf{n}_{\tau},\mathbf{n}_{\tau})\,d\mu_{\Sigma_{\tau}}= &\: 0,\\ \limsup_{\tau\to \infty}\int_{\Sigma_{\tau}\cap \{r\geq r_I\}} r^2|D_{r}\psi|^2\,\sin\theta d\theta d\varphi dr\geq &\: \mathfrak{c}_{\rm int}\sum_{\substack{\ell\in \mathbb{N}_0\\ \ell+\frac{1}{2}<|\mathfrak{q}Q|}}\sum_{\substack{m\in \mathbb{Z}\\ |m|\leq \ell}}|\mathfrak{I}_{\ell m}[\psi]|^2, \end{aligned}

   with induced volume form d\mu_{\tau} on \Sigma_{\tau} and \mathbf{n}_{\tau} the corresponding future-directed normal.

In the case that 1-\frac{|Q|}{M}\gtrsim 1 and |\mathfrak{q}Q|\gtrsim 1, the above results remain valid under the condition of the validity of mode stability on the real axis.

Applying standard “Big-O notation” (see for example [gaj26a][§ 2.2]), we can estimate:

Let \psi be a solution to (?) with A=\widehat{A}. Then:

With respect to the coordinate chart (\tau,\rho_{\infty},\theta,\varphi) we moreover obtain:

and with respect to the coordinate chart (\tau,\rho_+,\theta,\varphi), we obtain:

In particular, when \kappa_+=0, then

Note the symmetry between (2.2) and (2.2): we can obtain the expressions on the very RHS of (2.2) from the expressions to the right of the first equality in (2.2) by interchanging \rho_+ and \rho_{\infty}, replacing q with -q, T with K, \mathbbm{h} with \mathbbm{h}_+ and \widetilde{\mathbbm{h}} with \widetilde{\mathbbm{h}}_+.

1. (Poincaré inequality on \mathbb{S}^2) Let f\in H^1(\mathbb{S}^2), then:

   ||f_{\geq \ell}||_{L^2(\mathbb{S}^2)}\leq \frac{1}{\sqrt{\ell(\ell+1)}}||f||_{H^1(\mathbb{S}^2)}.

2. (Sobolev inequality on \mathbb{S}^2) Let f\in H^2(\mathbb{S}^2). Then there exists a numerical constant C>0, such that:

   ||f||_{L^{\infty}(\mathbb{S}^2)}\leq C ||f||_{H^2(\mathbb{S}^2)}\leq C\sum_{\ell\in \mathbb{N}_0}(\ell+1)^2||f_{\ell}||_{L^2(\mathbb{S}^2)}.

(Existence and uniqueness for the charged scalar field equation; [gaj26a][Theorem 2.5])

Let A\in \Omega^1(\mathcal{M}_{M,Q}) and G_A \in C^{\infty}(\mathcal{M}_{M,Q}). Consider the initial data pair:

Then there exists a unique solution \phi\in C^{\infty}(\mathcal{M}_{M,Q}) to (?) with respect to A, such that (\phi|_{\Sigma_0}, T\phi|_{\Sigma_0})=(\upphi,T\upphi).

If we take A\in \Omega^1(\widehat{\mathcal{M}}_{M,Q}) and G_A\in C^{\infty}(\widehat{\mathcal{M}}_{M,Q}), then \psi=r\phi\in C^{\infty}(\widehat{\mathcal{M}_{M,Q}}).

Let \psi be a solution to (?). Let \epsilon,\delta,\eta ,\kappa_0>0, N,\ell\in \mathbb{N}_0 and r_+<r_H<r_I<\infty. Consider 1< p<\min\{1+\textnormal{Re}\, \beta_{\ell},2\}+\epsilon and let \upzeta be a smooth function satisfying \upzeta(r)=1 for r\notin (r_{\sharp}-2\eta,r_{\sharp}+2\eta) and \upzeta(r)=0 for r\in (r_{\sharp}-\eta,r_{\sharp}+\eta).

Assume that 0\leq \kappa_+\leq \kappa_1 or |q|\leq q_1. Then, for suitably small \kappa_1 or q_1, there exist a constant
C=C(q_1,\epsilon,\delta,\eta, \kappa_1, p, r_H,r_I,\mathbbm{h},N)>0, such that for any \tau_1\leq \tau_2:

The estimate (?) remains valid when \kappa_+>\kappa_1 and |q|\geq q_1 under the assumption of Condition 1 below.

Let \ell\in \mathbb{N}_0, m\in \mathbb{Z} with |m|\leq \ell. Fix r_+=1. Let \mathfrak{w}_{\ell }, \mathfrak{w}^{\pm}_{\ell }: [r_+,\infty)\to \mathbb{C}, such that \mathfrak{w}_{\ell }Y_{\ell m} and \mathfrak{w}^{\pm}_{\ell }Y_{\ell m} are solutions to (?) with A=\widehat{A} and G_A=0. We write \mathfrak{w}^{\pm}(r;q) and \mathfrak{w}(r;q) when we need to emphasize the value of the corresponding charge parameter q.

Then:

1. We can fix \mathfrak{w}_{\ell }, \mathfrak{w}^{\pm}_{\ell } uniquely by setting:

   \begin{aligned} \mathfrak{w}_{\ell }(r_+)=&\:1,\\ \mathfrak{w}^{-}_{\ell }(r)=&\:r^{\frac{1}{2}+iq- \frac{1}{2}\beta_{\ell}}(1+O(r^{-1}))\quad (r>2),\\ \mathfrak{w}^{+}_{\ell }(r)=&\:r^{\frac{1}{2}+iq+ \frac{1}{2}\beta_{\ell}}(1+O(r^{-1}))\quad (\beta_{\ell}\neq 0)\quad (r>2),\\ \mathfrak{w}^{+}_{\ell }(r)=&\:r^{\frac{1}{2}+iq}\log r(1+O(r^{-1}))\quad (\beta_{\ell}= 0)\quad (r>2). \end{aligned}

2. We can express:

   \mathfrak{w}_{\ell }={\alpha}_+\mathfrak{w}_{\ell }^++{\alpha}_-\mathfrak{w}_{\ell }^-,

   with {\alpha}_{\pm}\in \mathbb{C}, such that in the case \beta_{\ell}\in i(0,\infty):

   \begin{gathered} \,\textnormal{sign}\,(q)\log\left|\frac{ \alpha_-}{\alpha_+}\right|=-\frac{\pi}{2}\left(\textnormal{Im}\, \beta_{\ell}-\frac{1}{\pi}\log \left(\frac{\cosh (\pi(|q|-\frac{1}{2}\textnormal{Im}\, \beta_{\ell}))}{\cosh (\pi(|q|+\frac{1}{2}\textnormal{Im}\, \beta_{\ell}))}\right)\right)\\ +\,\textnormal{sign}\,(q)\frac{\pi}{2}\left( \textnormal{Im}\, \beta_{\ell}-\frac{1}{\pi}\log \left(\frac{\cosh(\pi(\frac{|q|}{2}\frac{r_-+r_+}{r_+^2\kappa_+}+\frac{1}{2}\textnormal{Im}\, \beta_{\ell}))\cosh^2(\pi(|q|-\frac{1}{2}\textnormal{Im}\, \beta_{\ell}))}{\cosh(\pi(\frac{|q|}{2}\frac{r_-+r_+}{r_+^2\kappa_+}-\frac{1}{2}\textnormal{Im}\, \beta_{\ell}))\cosh^2(\pi(|q|+\frac{1}{2}\textnormal{Im}\, \beta_{\ell}))}\right)\right)\\ \substack{\kappa_+\downarrow 0\\ \to}-\frac{\pi}{2}\left(\textnormal{Im}\, \beta_{\ell}+\frac{1}{\pi}\log \left(\frac{\cosh (\pi(|q|-\frac{1}{2}\textnormal{Im}\, \beta_{\ell}))}{\cosh (\pi(|q|+\frac{1}{2}\textnormal{Im}\, \beta_{\ell}))}\right)\right). \end{gathered}

3. Denote with \alpha_{\pm}(\beta_{\ell}) the values of \alpha_{\pm} when \beta_{\ell}\neq 0 and q is fixed. Then \alpha_-(\beta_{\ell})=\alpha_+(-\beta_{\ell}) and we have that:

   \lim_{\beta_{\ell}\to 0}\mathfrak{w}_{\ell}(r)=\lim_{\beta_{\ell}\to 0}(\alpha_+\beta_{\ell})r^{\frac{1}{2}+iq}\left[\log r+\frac{d}{d\beta_{\ell}}\bigg|_{\beta_{\ell}=0}\frac{\alpha_{-}(\beta_{\ell})}{\alpha_{+}(\beta_{\ell})}\right].

   Therefore \alpha_{\pm} take the following values when \beta_{\ell}=0:

   \begin{aligned} \alpha_{+}=&\:-\lim_{\beta_{\ell}\to 0}(\alpha_+(\beta_{\ell})\beta_{\ell}),\\ \alpha_{-}=&\:\alpha_+ \frac{1}{i}\frac{d}{ds}\bigg|_{s=0}\frac{\alpha_{-}(is)}{\alpha_{+}(is)}. \end{aligned}

   In particular,

   \begin{gathered} \,\textnormal{sign}\,(q)\textnormal{Im}\, \left(\frac{\alpha_-}{\alpha_+}\right)=-\,\textnormal{sign}\,(q)\textnormal{Re}\,\left(\frac{d}{ds}\Big|_{s=0} \left(\frac{\alpha_-(is)}{\alpha_+(is)}\right)\right)=-\,\textnormal{sign}\,(q)\frac{ \pi}{2}(1-\tanh(\pi|q|))\\ +\frac{\pi}{2}\,\textnormal{sign}\,(q)\left(1 -\tanh\left(\pi \frac{|q|}{2}\frac{r_-+r_+}{r_+^2 \kappa_+}\right)-2\tanh(\pi |q|)\right)\\ \substack{\kappa_+\downarrow 0\\ \to} -\,\textnormal{sign}\,(q)\frac{ \pi}{2}(1+\tanh(\pi|q|). \end{gathered}

4. The function \mathfrak{w}_{\ell } is non-vanishing.

When \beta_{\ell}\in i[0,\infty), the following estimates hold for all \kappa_+\geq 0: there exists an \eta_0>0, such that:

Let \psi be a solution to (?) with \mathfrak{q}\neq 0 and
A=\widehat{A}=-Qr^{-1}d\tau, and write q=\mathfrak{q}Q. Let r_H>r_+. Let \beta_{\ell}=\sqrt{(2\ell+1)^2-4q^2}. If \kappa_+>\kappa_1 and |q|>q_1, assume moreover that Condition 1 holds.

1. Let N_1,N_2\in \mathbb{N}_0, l\in \{0,1\} and \nu,\epsilon>0. For arbitrary small \nu>0, there exists a constant C=C(\nu,\epsilon, N_1,N_2)>0, such that:

   \begin{aligned} \sum_{|\gamma|\leq N_2}||D_T^l\mathbf{Z}^{\gamma}(TK)^{N_1}(\psi-\Psi)||_{L^{\infty}(S^2_{\tau,r})}\\ \leq C \sum_{k\leq 2}\sqrt{E_{N_1+1,N_2+1+l,\epsilon}[\slashed{\nabla}_{\mathbb{S}^2}^k\psi]}(\tau+1)^{-1+\frac{1}{2}\nu-N_1-\frac{l}{2}}(r^{-1}\Omega\tau+1)^{-\frac{1}{2}-\frac{1}{2}\textnormal{Re}\, \beta_{0}},\\ \sum_{|\gamma|\leq N_2}||X^l\mathbf{Z}^{\gamma}(TK)^{N_1}(\psi-\Psi)||_{L^{\infty}(S^2_{\tau,r})}\\ \leq C\kappa_+^{-\frac{1}{2}} \sum_{k\leq 2}\sqrt{E_{N_1+1,N_2+1+l,\epsilon}[\slashed{\nabla}_{\mathbb{S}^2}^k\psi]}(\tau+1)^{-1+\frac{1}{2}\nu-N_1-\frac{l}{2}}(r^{-1}\tau+1)^{-\frac{1}{2}-\frac{1}{2}\textnormal{Re}\, \beta_{0}}\\ +Ce^{-c\kappa_+\tau}\sum_{|\gamma|\leq N_2}\sqrt{\int_{\Sigma_{0}\cap \{r\leq r_H\}}\mathcal{E}_2[\slashed{\nabla}_{\mathbb{S}^2}^kX^l\mathbf{Z}^{\gamma}(TK)^{N_1}\psi]\,d\sigma dr}\quad (|Q|<M), \end{aligned}

   with \Psi=\Psi^{\infty}+\Psi^{+}, where

   \begin{aligned} \Psi^{\infty}(\tau,r,\theta,\varphi):=&\:\sum_{\substack{\ell\in\mathbb{N}_0\\ \ell(\ell+1)<q^2}}\frac{\mathfrak{w}_{\ell}(r;q)}{\mathfrak{w}_{\ell}^0(r;q)}\sum_{\substack{m\in \mathbb{Z}\\ |m|\leq \ell}}\mathfrak{I}_{\ell m}[\psi](\Psi^{\infty}_0)_{\ell m}(\tau,r;q)Y_{\ell m}(\theta,\varphi),\\ \Psi^{+}(\tau,r,\theta,\varphi):=&\:e^{-iq r_+^{-1}\tau}\sum_{\substack{\ell\in\mathbb{N}_0\\ \ell(\ell+1)<q^2}}\frac{\mathfrak{w}_{\ell}\left(r_++\frac{r_+^2}{r-r_+};-q\right)}{\mathfrak{w}_{\ell}^0\left(r_++\frac{r_+^2}{r-r_+};-q\right)}\\ \times &\: \sum_{\substack{m\in \mathbb{Z}\\ |m|\leq \ell}}\mathfrak{H}_{\ell m}[\psi](\Psi^{\infty}_0)_{\ell m}\left(\tau,r_++\frac{r_+^2}{r-r_+};-q\right)Y_{\ell m}(\theta,\varphi), \end{aligned}

   where:

   • \mathfrak{I}_{\ell m}[\psi] and \mathfrak{h}_{\ell m}[\psi] are integrals that are linear functionals of derivatives of initial data (\psi,\partial_{\tau}\psi)|_{\tau=0} and \mathfrak{h}_{\ell m}[\psi]=0 for all data if |Q|<M; see the precise definitions in Proposition 8.1 and Corollary 8.2,

   • The constants \mathfrak{I}_{\ell m}[\psi] and, in the case |Q|=M, also \mathfrak{H}_{\ell m}[\psi], each vanish only for a codimension-1 subset of initial data,

   • the constants E_{N_1+1,N_2+1+l,\epsilon}[\slashed{\nabla}^k_{\mathbb{S}^2}\psi] are weighted, initial data energy norms that are defined in Proposition 9.1,

   • the functions \mathfrak{w}_{\ell,\pm }Y_{\ell m} are \tau-independent solutions to (?) with A=\widehat{A}, satisfying \mathfrak{w}_{\ell}=\alpha_+ \mathfrak{w}_{\ell}^{+}+\alpha_- \mathfrak{w}_{\ell}^{-}, with

     \begin{aligned} \mathfrak{w}_{\ell}^{-}(r;q)=&\:r^{\frac{1}{2}+iq- \frac{1}{2}\beta_{\ell}}(1+O(r^{-1}))\quad (r>2r_+),\\ \mathfrak{w}^{+}_{\ell }(r;q)=&\:r^{\frac{1}{2}+iq+ \frac{1}{2}\beta_{\ell}}(1+O(r^{-1}))\quad (\beta_{\ell}\neq 0)\quad (r>2r_+),\\ \mathfrak{w}^{+}_{\ell }(r;q)=&\:r^{\frac{1}{2}+iq}\log r(1+O(r^{-1}))\quad (\beta_{\ell}= 0)\quad (r>2r_+), \end{aligned}

     and \alpha_{\pm}(q)\in \mathbb{C} are constants such that \mathfrak{w}_{\ell}(r;q)={\alpha}_+(q) \mathfrak{w}_{\ell}^{+}(r;q)+{\alpha}_-(q) \mathfrak{w}_{\ell}^{-}(r;q) satisfies \mathfrak{w}_{\ell}(r_+;q)=1 (see also Proposition 3.1),

   • the functions \mathfrak{w}_{\ell,\pm }^0Y_{\ell m} are \tau-independent solutions to (?) with A=\widehat{A} and \mathbbm{h}=0, satisfying:

     \mathfrak{w}_{\ell}=\begin{cases} \alpha_+r^{\frac{1}{2}+iq+ \frac{1}{2}\beta_{\ell}}+\alpha_-r^{\frac{1}{2}+iq- \frac{1}{2}\beta_{\ell}}\quad &(\beta_{\ell}\neq 0),\\ \alpha_+r^{\frac{1}{2}+iq}\log r+\alpha_-r^{\frac{1}{2}+iq} &(\beta_{\ell}= 0). \end{cases}

   • The functions (\Psi_0)^{\infty}_{\ell m}Y_{\ell m}, with \ell(\ell+1)<q^2, are solutions to (?) with A=\widehat{A}, with:

     (\Psi^{\infty}_0)_{\ell m}(\tau,r;q):= \begin{cases} \beta_{\ell}\sum_{n=0}^{N_{\beta_{\ell}}}\zeta^n(1+\tau)^{-\frac{1}{2}-(n+\frac{1}{2})\beta_{\ell}+iq}\\ \times {}_2\mathbf{F}_1\left(\frac{1}{2}+\frac{\beta_{\ell}}{2}+iq,\frac{1}{2}-\frac{\beta_{\ell}}{2}+iq,\frac{1}{2}+iq-\left(n+\frac{1}{2}\right)\beta_{\ell};-\frac{\tau+1}{2r}\right),\quad \textnormal{when}\: \beta_{\ell}\in (0,1),\\ \mbox{}\\ \,\textnormal{sign}\,(q)\beta_{\ell}\sum_{n=0}^{\infty}\zeta^{\,\textnormal{sign}\,(q)n}(1+\tau)^{-\frac{1}{2}-(n+\frac{1}{2})\,\textnormal{sign}\,(q)\beta_{\ell}+iq}\\ \times {}_2\mathbf{F}_1\left(\frac{1}{2}+\,\textnormal{sign}\,(q)\frac{\beta_{\ell}}{2}+iq,\frac{1}{2}-\,\textnormal{sign}\,(q)\frac{\beta_{\ell}}{2}+iq,\frac{1}{2}+iq-\left(n+\frac{1}{2}\right)\,\textnormal{sign}\,(q)\beta_{\ell};-\frac{\tau+1}{2r}\right),\\ \textnormal{when}\:\beta_{\ell}\in i(0,\infty),\\ \mbox{}\\ (\tau+1)^{-\frac{1}{2}+iq}\int_0^{\infty}(\tau+1)^{-ix}e^{-\eta x}{}_2\mathbf{F}_1\left(\frac{1}{2}+iq,\frac{1}{2}+iq,\frac{1}{2}+iq-ix;-\frac{\tau+1}{2r}\right)\,dx,\\ \textnormal{when}\:\beta_{\ell}=0, \end{cases}

     with N_{\beta_{\ell}}=\lceil (\textnormal{Re}\, \beta_{\ell})^{-1}\rceil\in \mathbb{N}_0, {}_2\mathbf{F}_1(a,b,c;\cdot ) denoting regularized Gauss hypergeometric functions, see for example [NIST:DLMF][§ 15.2] for a definition, and where \zeta,\eta\in \mathbb{C} are defined as follows:

     \begin{aligned} \zeta(q):=& -\frac{2^{\beta_{\ell}}\Gamma(-\beta_{\ell}+1)\Gamma(\frac{1}{2}+ \frac{\beta_{\ell}}{2}+iq)}{\Gamma(\beta_{\ell}+1) \Gamma(\frac{1}{2}- \frac{\beta_{\ell}}{2}+iq)}\frac{\alpha_-(q)}{\alpha_+(q)},\\ \eta(q):=&\: i \,\textnormal{sign}\,(q)\left[\frac{\alpha_-(q)}{\alpha_+(q)}-\log 2+2\gamma_{\rm Euler}-\frac{\Gamma'(\frac{1}{2}+iq)}{\Gamma(\frac{1}{2}+iq)}\right], \end{aligned}

     where \gamma_{\rm Euler} is the Euler–Mascheroni constant.

2. In particular, we obtain along \mathcal{I}^+:

   (\Psi|_{\mathcal{I}^+})_{\ell m}(\tau)=\mathfrak{I}_{\ell m}[\psi]\begin{cases}\beta_{\ell}(1+\tau)^{-\frac{1}{2}-\frac{\beta_{\ell}}{2}+iq}\sum_{n=0}^{N_{\beta_{\ell}}}\frac{\zeta^n(q)(1+\tau)^{-n\beta_{\ell}}}{\Gamma\left(\frac{1}{2}+iq-(n+\frac{1}{2})\beta_{\ell}\right)}\\ +e^{-iq r_+^{-1}\tau}O_{\infty}((\tau+1)^{-1-\beta_{\ell}}),\quad &(\beta_{\ell}\in (0,1)),\\ \,\textnormal{sign}\,(q)\beta_{\ell}(1+\tau)^{-\frac{1}{2}-\,\textnormal{sign}\,(q)\frac{\beta_{\ell}}{2}+iq}\sum_{n=0}^{\infty}\frac{\zeta^{\,\textnormal{sign}\,(q)n}(q)(1+\tau)^{-\,\textnormal{sign}\,(q)n\beta_{\ell}}}{\Gamma\left(\frac{1}{2}+iq-(n+\frac{1}{2})\,\textnormal{sign}\,(q)\beta_{\ell}\right)}\\ +e^{-iq r_+^{-1}\tau}O_{\infty}((\tau+1)^{-1}),\quad &(\beta_{\ell}\in i(0,\infty)),\\ \frac{-i}{\Gamma\left(\frac{1}{2}+iq\right)}(\tau+1)^{-\frac{1}{2}+iq}\left[(\log(1+\tau))^{-1}+O((\log(1+\tau))^{-2})\right]\\ +e^{-iq r_+^{-1}\tau}\frac{1}{\log(1+\tau)}O_{\infty}((\tau+1)^{-1}),\quad &(\beta_{\ell}=0). \end{cases}

   Let r_0>r_+. Then we obtain along \{r=r_0\}:

   \begin{aligned} (\Psi)_{\ell m}(\tau,r_0)=&-2^{\frac{1}{2}+\frac{1}{2}\beta_{\ell}}\mathfrak{I}_{\ell m}[\psi]\frac{\pi \beta_{\ell}}{\alpha_+(q)\sin (\pi \beta_{\ell})}(1+\tau)^{-1-\beta_{\ell}}\mathfrak{w}_{\ell}(r_0;q)\\ \times & \left[\sum_{n=0}^{N_{\beta_{\ell}}}\zeta^n(q)\frac{(1+\tau)^{-n\beta_{\ell}}}{\Gamma(\beta_{\ell}+1) \Gamma(\frac{1}{2}-\frac{1}{2}\beta_{\ell}+iq)\Gamma(-(n+1)\beta_{\ell})}\right] \\ &-2^{\frac{1}{2}+\frac{1}{2}\beta_{\ell}}\mathfrak{H}_{\ell m}[\psi]\frac{\pi \beta_{\ell}}{\alpha_+(-q)\sin (\pi \beta_{\ell})}(1+\tau)^{-1-\beta_{\ell}}e^{-iq r_+^{-1}\tau}\mathfrak{w}_{\ell}(r_++r_+^{2}(r_0-r_+)^{-1};-q)\\ \times &\left[\sum_{n=0}^{N_{\beta_{\ell}}}\zeta^n(-q)(\beta_{\ell}) \frac{(1+\tau)^{-n\beta_{\ell}}}{\Gamma(\beta_{\ell}+1) \Gamma(\frac{1}{2}-\frac{1}{2}\beta_{\ell}-iq)\Gamma(-(n+1)\beta_{\ell})}\right]\\ &+O_{\infty}\left((1+\tau)^{-2}\right)+e^{-iq r_+^{-1}\tau}O_{\infty}\left((1+\tau)^{-2}\right), \quad (\beta_{\ell}\in (0,1)),\\ (\Psi)_{\ell m}(\tau,r_0)=&-2^{\frac{1}{2}+\frac{1}{2}\,\textnormal{sign}\,(q)\beta_{\ell}}\mathfrak{I}_{\ell m}[\psi]\frac{\pi \beta_{\ell}}{\alpha_+(q)\sin (\pi \beta_{\ell})}(1+\tau)^{-1-\,\textnormal{sign}\,(q)\beta_{\ell}}\mathfrak{w}_{\ell}(r_0;q)\\ \times & \left[\sum_{n=0}^{\infty}\zeta^{\,\textnormal{sign}\,(q)n}(q) \frac{(1+\tau)^{-\,\textnormal{sign}\,(q)n\beta_{\ell}}}{\Gamma(\,\textnormal{sign}\,(q)\beta_{\ell}+1) \Gamma(\frac{1}{2}-\frac{1}{2}\,\textnormal{sign}\,(q)\beta_{\ell}+iq)\Gamma(-\,\textnormal{sign}\,(q)(n+1)\beta_{\ell})}\right] \\ &-2^{\frac{1}{2}-\frac{1}{2}\,\textnormal{sign}\,(q)\beta_{\ell}}\mathfrak{H}_{\ell m}[\psi]\frac{\pi \beta_{\ell}}{\alpha_+(-q)\sin (\pi \beta_{\ell})}(1+\tau)^{-1+\,\textnormal{sign}\,(q)\beta_{\ell}}e^{-iq r_+^{-1}\tau}\mathfrak{w}_{\ell}(r_++r_+^{2}(r_0-r_+)^{-1};–q)\\ \times& \left[\sum_{n=0}^{\infty}\zeta^{-\,\textnormal{sign}\,(q)n}(-q) \frac{(1+\tau)^{\,\textnormal{sign}\,(q)n\beta_{\ell}}}{\Gamma(-\,\textnormal{sign}\,(q)\beta_{\ell}+1) \Gamma(\frac{1}{2}+\frac{1}{2}\,\textnormal{sign}\,(q)\beta_{\ell}-iq)\Gamma((n+1)\,\textnormal{sign}\,(q)\beta_{\ell})}\right]\\ &+O_{\infty}\left((1+\tau)^{-2}\right)+e^{-iq r_+^{-1}\tau}O_{\infty}\left((1+\tau)^{-2}\right), \quad (\beta_{\ell}\in i(0,\infty)),\\ (\Psi)_{\ell m}(\tau,r_0)=&-\frac{2^{\frac{1}{2}+iq}}{\Gamma(\frac{1}{2}+iq)}\mathfrak{I}_{\ell m}[\psi]\mathfrak{w}_{\ell}(r_0;q)\left[(\log(1+\tau))^{-2}(\tau+1)^{-1}+(\log(1+\tau))^{-3}O_{\infty}((\tau+1)^{-1}\right]\\ &-\frac{2^{\frac{1}{2}-iq}}{\Gamma(\frac{1}{2}-iq)}e^{-iq r_+^{-1}\tau}\mathfrak{H}_{\ell m}[\psi]\mathfrak{w}_{\ell}(r_++r_+^{2}(r_0-r_+)^{-1};-q)\\ &\times \left[(\log(1+\tau))^{-2}(\tau+1)^{-1}+(\log(1+\tau))^{-3}O_{\infty}((\tau+1)^{-1}\right]\\ &+O_{\infty}((1+\tau)^{-2})+e^{-iq r_+^{-1}\tau}O_{\infty}((1+\tau)^{-2}), \quad (\beta_{\ell}=0). \end{aligned}

   We obtain along \mathcal{H}^+:

   \begin{aligned} (\Psi)_{\ell m}(\tau,r_+)=&\:\beta_{\ell}\mathfrak{H}_{\ell m}[\psi]e^{-iq r_+^{-1}\tau}(1+\tau)^{-\frac{1}{2}-\frac{\beta_{\ell}}{2}-iq} \sum_{n=0}^{N_{\beta_{\ell}}}\frac{\zeta^n(-q)(1+\tau)^{-n\beta_{\ell}}}{\Gamma\left(\frac{1}{2}-iq-(n+\frac{1}{2})\beta_{\ell}\right)}\\ &-2^{\frac{1}{2}+\frac{1}{2}\beta_{\ell}}\mathfrak{I}_{\ell m}[\psi]\frac{\pi \beta_{\ell}}{\alpha_+(q)\sin (\pi \beta_{\ell})}(1+\tau)^{-1-\beta_{\ell}}\\ \times & \left[\sum_{n=0}^{N_{\beta_{\ell}}}\zeta^n(q)\frac{(1+\tau)^{-n\beta_{\ell}}}{\Gamma(\beta_{\ell}+1) \Gamma(\frac{1}{2}-\frac{1}{2}\beta_{\ell}+iq)\Gamma(-(n+1)\beta_{\ell})}\right] \\ &+O_{\infty}\left((1+\tau)^{-2}\right), \quad (\beta_{\ell}\in (0,1)),\\ (\Psi)_{\ell m}(\tau,r_+)=&\:-\,\textnormal{sign}\,(q)\beta_{\ell}\mathfrak{H}_{\ell m}[\psi]e^{-iq r_+^{-1}\tau}(1+\tau)^{-\frac{1}{2}+\,\textnormal{sign}\,(q)\frac{\beta_{\ell}}{2}-iq}\sum_{n=0}^{\infty}\frac{\zeta^{-\,\textnormal{sign}\,(q)n}(-q)(1+\tau)^{\,\textnormal{sign}\,(q)n\beta_{\ell}}}{\Gamma\left(\frac{1}{2}-iq+(n+\frac{1}{2})\,\textnormal{sign}\,(q)\beta_{\ell}\right)}\\ &-2^{\frac{1}{2}+\frac{1}{2}\,\textnormal{sign}\,(q)\beta_{\ell}}\mathfrak{I}_{\ell m}[\psi]\frac{\pi \beta_{\ell}}{\alpha_+(q)\sin (\pi \beta_{\ell})}(1+\tau)^{-1-\,\textnormal{sign}\,(q)\beta_{\ell}}\\ \times & \left[\sum_{n=0}^{\infty}\zeta^{\,\textnormal{sign}\,(q)n}(q) \frac{(1+\tau)^{-\,\textnormal{sign}\,(q)n\beta_{\ell}}}{\Gamma(\,\textnormal{sign}\,(q)\beta_{\ell}+1) \Gamma(\frac{1}{2}-\frac{1}{2}\,\textnormal{sign}\,(q)\beta_{\ell}+iq)\Gamma(-\,\textnormal{sign}\,(q)(n+1)\beta_{\ell})}\right] \\ &+O_{\infty}\left((1+\tau)^{-2}\right), \quad (\beta_{\ell}\in i(0,\infty)),\\ (\Psi)_{\ell m}(\tau,r_+)=&\: \frac{-i}{\Gamma\left(\frac{1}{2}-iq\right)}\mathfrak{H}_{\ell m}[\psi]e^{-iq r_+^{-1}\tau}(\tau+1)^{-\frac{1}{2}-iq}\left[(\log(1+\tau))^{-1}+O((\log(1+\tau))^{-2})\right]\\ &-\frac{2^{\frac{1}{2}+iq}}{\Gamma(\frac{1}{2}+iq)}\mathfrak{I}_{\ell m}[\psi]\left[(\log(1+\tau))^{-2}(\tau+1)^{-1}+(\log(1+\tau))^{-3}O_{\infty}((\tau+1)^{-1}\right]\\ &+O_{\infty}((1+\tau)^{-2}),\quad (\beta_{\ell}=0). \end{aligned}

3. We obtain moreover the following late-time asymptotics for D_T\Psi along \mathcal{H}^+ and \mathcal{I}^+:

   \begin{aligned} (D_T\Psi|_{\mathcal{I}^+})_{\ell m}(\tau)=&\:2^{\frac{1}{2}+\frac{1}{2}\beta_{\ell}}q r_+^{-1}\mathfrak{H}_{\ell m}[\psi]\frac{\pi \beta_{\ell}}{\alpha_+(-q)\sin (\pi \beta_{\ell})}(1+\tau)^{-1-\beta_{\ell}}e^{-iq r_+^{-1}\tau}\mathfrak{w}_{\ell}(r_+;-q)\\ \times &\left[\sum_{n=0}^{N_{\beta_{\ell}}}\zeta^n(-q)(\beta_{\ell}) \frac{(1+\tau)^{-n\beta_{\ell}}}{\Gamma(\beta_{\ell}+1) \Gamma(\frac{1}{2}-\frac{1}{2}\beta_{\ell}-iq)\Gamma(-(n+1)\beta_{\ell})}\right]\\ +&\:O_{\infty}((\tau+r_+)^{-\frac{3+\beta_{\ell}}{2} }), \quad (\beta_{\ell}\in (0,1),\\ (D_T\Psi|_{\mathcal{I}^+})_{\ell m}(\tau)=&\:2^{\frac{1}{2}-\frac{1}{2}\,\textnormal{sign}\,(q)\beta_{\ell}}\mathfrak{H}_{\ell m}[\psi]q r_+^{-1}\frac{\pi \beta_{\ell}}{\alpha_+(-q)\sin (\pi \beta_{\ell})}(1+\tau)^{-1+\,\textnormal{sign}\,(q)\beta_{\ell}}e^{-iq r_+^{-1}\tau}\mathfrak{w}_{\ell}(r_+;–q)\\ \times & \left[\sum_{n=0}^{\infty}\zeta^{-\,\textnormal{sign}\,(q)n}(-q) \frac{(1+\tau)^{-\,\textnormal{sign}\,(q)n\beta_{\ell}}}{\Gamma(-\,\textnormal{sign}\,(q)\beta_{\ell}+1) \Gamma(\frac{1}{2}+\frac{1}{2}\,\textnormal{sign}\,(q)\beta_{\ell}-iq)\Gamma((n+1)\,\textnormal{sign}\,(q)\beta_{\ell})}\right]\\ +&\:O_{\infty}((\tau+r_+)^{-\frac{3}{2} }), \quad (\beta_{\ell}\in i(0,\infty)),\\ (D_T\Psi|_{\mathcal{I}^+})_{\ell m}(\tau)=&\:\frac{2^{\frac{1}{2}-iq}}{\Gamma(\frac{1}{2}-iq)}q r_+^{-1}e^{-iq r_+^{-1}\tau}\mathfrak{H}_{\ell m}[\psi]\mathfrak{w}_{\ell}(r_++r_+^{2}(r_0-r_+)^{-1};-q)\\ &\times \left[(\log(1+\tau))^{-2}(\tau+1)^{-1}+(\log(1+\tau))^{-3}O_{\infty}((\tau+1)^{-1}\right]\\ +&\:O_{\infty}((\tau+r_+)^{-\frac{3}{2} }), \quad (\beta_{\ell}=0),\\ (D_T\Psi|_{\mathcal{H}^+})_{\ell m}(\tau)=&\:-2^{\frac{1}{2}-\frac{1}{2}\beta_{\ell}}q r_+^{-1}\mathfrak{I}_{\ell m}[\psi]\frac{\pi \beta_{\ell}}{\alpha_+(q)\sin (\pi \beta_{\ell})}(1+\tau)^{-1-\beta_{\ell}}\mathfrak{w}_{\ell}(r_+;q)\\ \times &\left[\sum_{n=0}^{N_{\beta_{\ell}}}\zeta^n(q)(\beta_{\ell}) \frac{(1+\tau)^{-n\beta_{\ell}}}{\Gamma(\beta_{\ell}+1) \Gamma(\frac{1}{2}-\frac{1}{2}\beta_{\ell}+iq)\Gamma(-(n+1)\beta_{\ell})}\right]\\ +&\:O_{\infty}((\tau+r_+)^{-\frac{3+\beta_{\ell}}{2} }), \quad (\beta_{\ell}\in (0,1),\\ (D_T\Psi|_{\mathcal{H}^+})_{\ell m}(\tau)=&\:-2^{\frac{1}{2}+\frac{1}{2}\,\textnormal{sign}\,(q)\beta_{\ell}}q r_+^{-1}\mathfrak{I}_{\ell m}[\psi]\frac{\pi \beta_{\ell}}{\alpha_+(q)\sin (\pi \beta_{\ell})}(1+\tau)^{-1-\,\textnormal{sign}\,(q)\beta_{\ell}}\mathfrak{w}_{\ell}(r_+;q)\\ \times& \left[\sum_{n=0}^{\infty}\zeta^{\,\textnormal{sign}\,(q)n}(q) \frac{(1+\tau)^{-\,\textnormal{sign}\,(q)n\beta_{\ell}}}{\Gamma(\,\textnormal{sign}\,(q)\beta_{\ell}+1) \Gamma(\frac{1}{2}-\frac{1}{2}\,\textnormal{sign}\,(q)\beta_{\ell}+iq)\Gamma(-(n+1)\,\textnormal{sign}\,(q)\beta_{\ell})}\right]\\ +&\:O_{\infty}((\tau+r_+)^{-\frac{3}{2} }), \quad (\beta_{\ell}\in i(0,\infty)),\\ (D_T\Psi|_{\mathcal{H}^+})_{\ell m}(\tau)=&\:-\frac{2^{\frac{1}{2}+iq}}{\Gamma(\frac{1}{2}+iq)}q r_+^{-1}\mathfrak{I}_{\ell m}[\psi]\mathfrak{w}_{\ell}(r_+;q)\\ &\times \left[(\log(1+\tau))^{-2}(\tau+1)^{-1}+(\log(1+\tau))^{-3}O_{\infty}((\tau+1)^{-1}\right]\\ +&\:O_{\infty}((\tau+r_+)^{-\frac{3}{2} }), \quad (\beta_{\ell}=0). \end{aligned}

Let \psi be as in Theorem 4.1. If \kappa_+>\kappa_1 and |q|>q_1, assume moreover that Condition 1 holds. Let N_1\in \mathbb{N}_0, N_2\in \mathbb{N}_0. Let 1< p<2. For all \nu>0 and \epsilon>0 suitably small, there exists a constant C=C(\nu,\epsilon, N_1,N_2,p)>0 such that for |\gamma|\leq N_2:

Furthermore, for p\in [0,1-\epsilon]

and

Let \psi be as in Theorem 4.1. If \kappa_+>\kappa_1 and |q|>q_1, assume moreover that Condition 1 holds. Let \nu,\epsilon,\delta>0 be arbitrarily small. Let \mathfrak{c}_k\in \mathbb{C} and \mathfrak{c}_{\rm int}\in (0,\infty) be the constants defined in Proposition A.7.

1. Let r_H>r_+ and r_I<\infty be arbitrary. Then for \kappa_+\leq \kappa_0:

   \begin{aligned} \limsup_{\tau\to \infty}(1+\tau)^{\textnormal{Re}\, \beta_0}\int_{\Sigma_{\tau}\cap\{r\leq r_H\}}\mathcal{E}_2[\psi]\,d\sigma dr\geq &\: \mathfrak{c}_{\rm int}\sum_{\substack{\ell\in \mathbb{N}_0\\ \textnormal{Re}\, \beta_{\ell}=\textnormal{Re}\, \beta_{0}}}\sum_{\substack{m\in \mathbb{Z}\\ |m|\leq \ell}}|\mathfrak{H}_{\ell m}[\psi]|^2\quad &( |Q|=M,\: \beta_{0}\neq 0),\\ \limsup_{\tau\to \infty}\log^2(1+\tau)\int_{\Sigma_{\tau}\cap\{r\leq r_H\}}\mathcal{E}_2[\psi]\,d\sigma dr\geq &\: \mathfrak{c}_{\rm int}|\mathfrak{H}_{00}[\psi]|^2\quad &( |Q|=M,\: \beta_{0}= 0),\\ \limsup_{\tau\to \infty}(1+\tau)^{\textnormal{Re}\, \beta_{0}}\int_{\Sigma_{\tau}\cap\{r\geq r_I\}}\mathcal{E}_2[\psi]\,d\sigma dr\geq &\: \mathfrak{c}_{\rm int}\sum_{\substack{\ell\in \mathbb{N}_0\\ \textnormal{Re}\, \beta_{\ell}=\textnormal{Re}\, \beta_{0}}}\sum_{\substack{m\in \mathbb{Z}\\ |m|\leq \ell}}|\mathfrak{I}_{\ell m}[\psi]|^2\quad &( \beta_{0}\neq 0),\\ \limsup_{\tau\to \infty}\log^2(1+\tau)\int_{\Sigma_{\tau}\cap\{r\geq r_I\}}\mathcal{E}_2[\psi]\,d\sigma dr\geq &\: \mathfrak{c}_{\rm int}|\mathfrak{I}_{00}[\psi]|^2\quad &( \beta_{0}= 0). \end{aligned}

   Furthermore, \limsup_{\tau\to \infty}\log^2(1+\tau)(1+\tau)^{\textnormal{Re}\, \beta_0}\int_{\Sigma_{\tau}\cap\{r_H\leq r\leq r_I\}}\mathcal{E}_2[\psi]\,d\sigma dr=0 for all \beta_0\in [0,1]\cup i(0,\infty).

2. Let k\in \mathbb{N}_0 and \ell(\ell+1)<q^2. Then there exists a constant C_k>0, such that for all

   \begin{aligned} \limsup_{\tau\to \infty}(1+\tau)^{-k-\frac{1}{2}+\frac{\textnormal{Re}\, \beta_{\ell}}{2}}|(r^2X)^{k+1}\psi_{\ell m}|(\tau,r_+)\geq &\:|\mathfrak{c}_k| |\mathfrak{H}_{\ell m}[\psi]| \quad &(|Q|=M,\:\beta_{\ell}\neq 0),\\ \limsup_{\tau\to \infty}\log(1+\tau)(1+\tau)^{-k-\frac{1}{2}}|(r^2X)^{k+1}\psi_{\ell m}|(\tau,r_+)\geq &\:|\mathfrak{c}_k| |\mathfrak{H}_{\ell m}[\psi]|\quad &(|Q|=M,\:\beta_{\ell}=0),\\ \limsup_{\tau\to \infty}(1+\tau)^{-k-\frac{1}{2}+\frac{\textnormal{Re}\, \beta_{\ell}}{2}}|(r^2X)^{k+1}\psi_{\ell m}|(\tau,\infty)\geq &\:|\mathfrak{c}_k| |\mathfrak{I}_{\ell m}[\psi]| \quad &(\beta_{\ell}\neq 0),\\ \limsup_{\tau\to \infty}\log(1+\tau)(1+\tau)^{-k-\frac{1}{2}}|(r^2X)^{k+1}\psi_{\ell m}|(\tau,\infty)\geq &\:|\mathfrak{c}_k| |\mathfrak{I}_{\ell m}[\psi]|\quad &(\beta_{\ell}=0). \end{aligned}

Let N\in \mathbb{N}_0.

1. Let \psi be a solution to (?). Denote \psi^{(N)}=(\rho_{\infty}\partial_{\rho_{\infty}})^N\psi. Then:

   \begin{gathered} \rho_{\infty}^{-1}(\rho_{\infty}\partial_{\rho_{\infty}})^N(\rho_{\infty} r^3G_{\widehat{A}})=\partial_{\rho_{\infty}}(\Omega^2r^{-2}\partial_{\rho_{\infty}}\psi^{(N)})+\slashed{\Delta}_{\mathbb{S}^2}\psi^{(N)}+2\partial_{\rho_{\infty}}T\psi^{(N)}\\ +2iq\rho_{\infty} \partial_{\rho_{\infty}}\psi^{(N)}+N(1+O_{\infty}(\rho_{\infty}))\rho_{\infty}\partial_{\rho_{\infty}}\psi^{(N)}+N \slashed{\Delta}_{\mathbb{S}^2}\psi^{(N-1)}\\ -\left[r\frac{d\Omega^2}{dr}-i q(1+2N) -\frac{1}{2}N(N+1)+O_{\infty}(\rho_{\infty})\right]\psi^{(N)}+N\sum_{n=0}^{N-1} O_{\infty}(\rho_{\infty}^0)\psi^{(n)}\\ +N(N-1)\sum_{n=0}^{N-2} O_{\infty}(\rho_{\infty}^0)\slashed{\Delta}_{\mathbb{S}^2}\psi^{(n)} \end{gathered}

2. Let \psi be a solution to (?). Denote \psi^{(N)}=(\rho_+\partial_{\rho_+})^N\psi. Then:

   \begin{gathered} \rho_+^{-1}(\rho_+\partial_{\rho_+})^N(\rho_+ r^3G_{\widehat{A}})=\partial_{\rho_+}(\Omega^2r^{-2}\partial_{\rho_+}\psi^{(N)})+\slashed{\Delta}_{\mathbb{S}^2}\psi^{(N)}+2\partial_{\rho_+}K\psi^{(N)}\\ - 2iq\rho_+ \partial_{\rho_+}\psi^{(N)}+N(1-6\kappa_+ r_+)(1+O_{\infty}(\rho_+))\rho_+\partial_{\rho_+}\psi^{(N)}+N \slashed{\Delta}_{\mathbb{S}^2}\psi^{(N-1)}\\ -\left[r\frac{d\Omega^2}{dr}+i q(1+2N) -\frac{1}{2}N(N+1)(1-6\kappa_+ r_+)+O_{\infty}(\rho_+)\right]\psi^{(N)}+N\sum_{n=0}^{N-1} O_{\infty}(\rho_+^0)\psi^{(n)}\\ +N(N-1)\sum_{n=0}^{N-2} O_{\infty}(\rho_+^0)\slashed{\Delta}_{\mathbb{S}^2}\psi^{(n)}. \end{gathered}

Let N\in \mathbb{N}_0 and let \psi be a solution to (?). Let 1<r_H<r_I<\infty, q\in \mathbb{R}, \epsilon>0, K\in \mathbb{N}_0 and 1< p<\min\{1+\textnormal{Re}\, \beta_{\ell},2\}+\epsilon. Then there exist a constant C=C(\mathbbm{h},\epsilon,p, r_H,r_I,N)>0, such that for suitably small \epsilon>0:

Let \gamma\in \mathbb{N}_0 and \epsilon>0. Let 1< p<\min\{1+\textnormal{Re}\, \beta_{\ell},2\}+\epsilon. Then

Let N_1\in \mathbb{N}_1, N_2\in \mathbb{N}_0 and q_1\in [0,\infty). Let \psi be a solution to (?) with A=\widehat{A} and |q|\leq q_1. For all \nu>0, there exists an \epsilon_0>0, such that for all 0\leq \epsilon\leq \epsilon_0 and 1+\epsilon\leq p\leq 1+\textnormal{Re}\,\sqrt{1-4q_1^2}+\epsilon, there exists a constant C=C(\mathbbm{h},\nu,\epsilon_0,p, N_1,N_2,q_1)>0 such that:

Let N_1\in \mathbb{N}_1, N_2\in \mathbb{N}_0 and q_1\in [0,\infty). Let |q|\leq q_1. Let 1+\textnormal{Re}\, \sqrt{1-4q_1^2}< p<2+\textnormal{Re}\, \sqrt{1-4q_1^2}. For all \nu>0 and 0<\epsilon<p-1-\textnormal{Re}\, \sqrt{1-4q_1^2}<1, there exists a constant
C=C(\mathbbm{h},\nu,\epsilon, N_1,N_2,p,q_1)>0 such that:

Fix r_+=1. Let n_1,n_2\in \mathbb{N}_0. Then there exists a constant C=C(n_1,n_2,\widetilde{\mathbbm{h}})>0 such that:

and

Let r_+=1 and s_+=(r-1)^{-1}. The following constants are well-defined:

Fix r_+=1. Suppose that B_{\infty,\ell m}\neq 0 and fix the constants \mathfrak{I}_{\ell m}[\psi] appearing in the definition of \Psi^{\infty} as follows:

Then there exists a unique solution T^{-1}(\psi_{\ell m}-\Psi^{\infty}_{\ell m})|_{\Sigma_0}\in C^{\infty}([r_+,\infty)) to (?) such that

We denote with

the solution to (?) with A=\widehat{A} and G=G[\Psi^{\infty}], arising from the initial data:

Then T(T^{-1}(\psi_{\ell}-\Psi^{\infty}_{\ell}))=\psi_{\ell}-\Psi^{\infty}_{\ell}.

Let |Q|=M. Fix r_+=1. Suppose that B_{+,\ell m}\neq 0 and fix:

Then there exists a unique solution K^{-1}(\psi_{\ell}-\Psi^{+}_{\ell})|_{\Sigma_0}\in C^{\infty}([0,r_+^{-1})_{\rho_{\infty}}) to (?) such that

We denote with

the solution to (?) with A=\widehat{A} and G=G[\Psi^{+}], arising from the initial data:

Then K(K^{-1}(\psi_{\ell}-\Psi^{\infty}_{\ell}))=\psi_{\ell}-\Psi^{+}_{\ell}.

Let L_0\in \mathbb{N}_0, \tau\geq 0 and let \phi be a solution to (?) with A=\widehat{A} and G_A\equiv 0. Let 1<p< 3. Then for suitable large L_0>0, depending on q and p, there exists a constant C=C(p,q,L_0)>0 such that for all \tau\geq 0:

Let L_0,N\in \mathbb{N}_0. Let \tau\geq 0 and let \phi be a solution to (?) with G\equiv 0.

Let 1<p<3. Then, for suitably large L_0>0 depending on q and p, there exists a constant C=C(p,q,L_0,N)>0 such that for all \tau\geq 0:

Let r>1, \delta>0 and N\in \mathbb{N}_0. Then

is well-defined as an element of the Sobolev space H^N(\mathbb{S}^2), provided

Let N,L_0\in \mathbb{N}_0. Let \tau\geq 0 and let \phi be a solution to (?). Let 2-\textnormal{Re}\, \sqrt{1-4q^2}<p<3 and \epsilon>0. Then there exists a constant C=C(p,q,L_0,\epsilon)>0 such that for all \tau\geq 0:

where (8.6) only holds for |Q|=M.

Let N_1\in \mathbb{N}_0, N_2\in \mathbb{N}_0. Let 1< p<2. For all \nu>0 and \epsilon>0 suitably small, there exists a constant C=C(\nu,\epsilon, N_1,N_2,p)>0 such that:

and additionally,

Furthermore, for N_1\in \mathbb{N}_0\cup\{-1\}, N_2\in \mathbb{N}_0:

Let N_1\in \mathbb{N}_0\cup\{-1\}, N_2\in \mathbb{N}_0. Let l\in\{0,1\}. For arbitrarily small \nu>0 and \epsilon>0, there exists a constant C=C(\nu,\epsilon, N_1,N_2,p)>0 such that:

If |Q|<M, we moreover have for all l\in \mathbb{N}_0:

B_{\infty,\ell m}\neq 0 and B_{+,\ell m}\neq 0 for all \ell\in \mathbb{N}_0 such that \ell(\ell+1)<q^2.

Let \psi be a solution to (?) with G_A=0 and A=-\frac{Q}{r}d\tau and \kappa_+>0. Let k\in \mathbb{N}_0. Fix the initial data for \psi independently of \kappa_+. Let \psi_M be a solution arising from the same data, but with \kappa_+=0. Then, there exists suitably small 0<\delta_1<\delta_2, depending only on the initial data for \psi and on k, such that for r_+\kappa_+\leq \delta_1 and \ell(\ell+1)<q^2:

1. Let s\in \mathbb{C}\setminus \mathbb{Z}. The function (\tau+1)^{s}F(\varrho)Y_{\ell m}(\theta,\varphi) is a solution to (?) with A=\widehat{A} and \mathbbm{h}=0 that is smooth in \varrho at \varrho=0 if and only if

   F(\varrho)= B\cdot {}_2\mathbf{F}_1(a,b,c;-\varrho),

   with {}_2\mathbf{F}_1(a,b,c;\cdot) the regularized Gauss hypergeometric function, B\in \mathbb{C} and

   \begin{aligned} a=&\:\frac{1}{2}\pm \frac{\beta_{\ell}}{2}+iq,\\ b=&\:\frac{1}{2}\mp \frac{\beta_{\ell}}{2}+iq,\\ c=&\: s+1. \end{aligned}

   In particular, we obtain the following simplifications for particular choices of s: let a=\frac{1}{2}+ \frac{\beta_{\ell}}{2}+iq, then

   \begin{aligned} (\tau+1)^sF(\varrho)=&\:B(1+\tau)^{-\frac{1}{2}-\frac{\beta_{\ell}}{2}+iq}(1+\varrho)^{-\frac{1}{2}-\frac{\beta_{\ell}}{2}-iq}\quad (s=b-1),\\ (\tau+1)^sF(\varrho)=&\:B(1+\tau)^{-\frac{1}{2}+\frac{\beta_{\ell}}{2}+iq}(1+\varrho)^{-\frac{1}{2}+\frac{\beta_{\ell}}{2}-iq}\quad (s=a-1). \end{aligned}

2. Let \beta_{\ell}=0 and define

   H(\tau,\varrho):=(\tau+1)^{-\frac{1}{2}+iq}(1+\varrho)^{-\frac{1}{2}-iq}(\log(\varrho+1)+\log (\tau+1)).

   Then H(\tau,\varrho)Y_{\ell m}(\theta,\varphi) is a solution to (?) with A=\widehat{A} and \mathbbm{h}=0 that is smooth in \varrho at \varrho=0.

Let \beta_{\ell}\in (0,1) and \alpha_+,\alpha_-\in \mathbb{C}. Let \mathfrak{w}_{\ell}^0(r)=\alpha_+ r^{\frac{1}{2}+\frac{\beta_{\ell}}{2}+iq}+\alpha_-r^{\frac{1}{2}-\frac{\beta_{\ell}}{2}+iq} and let r\geq R, so that \mathfrak{w}_0(r)\neq 0. Let N_{\beta_{\ell}}\in \mathbb{N}_0, such that N_{\beta_{\ell}}\beta_{\ell}>1.

Define

with

Then \Psi^{\infty}_0=(\Psi^{\infty}_0)_{\ell m}Y_{\ell m} is smooth with respect to the coordinate chart (\tau,\varrho_{\infty},\theta,\varphi) (where \varrho_{\infty}=\frac{1}{r}) and it is a solution to (?) with A=\widehat{A} and \mathbbm{h}=0.

Furthermore, f(\tau,r):=(\mathfrak{w}^0_{\ell})^{-1}(r)(\Psi^{\infty}_0)_{\ell m}(\tau,r) satisfies: for all k,l\leq N

Let \beta \in i[0,\infty), q\in \mathbb{R} and x\in (0,\infty). Define F_x(\varrho;\beta):={}_2\mathbf{F}_1(\frac{1}{2}+ \frac{\beta}{2}+iq,\frac{1}{2}- \frac{\beta}{2}+iq,\frac{1}{2}-\frac{\beta}{2}+i(q+x);-\varrho).

1. Let \varrho_0>0. Then for all N\in \mathbb{N}_0, there exists a constant C_{N,\varrho_0,q}>0, such that for all 0\leq m\leq N :

   \left|((\varrho\partial_{\varrho})^m \left(\Gamma\left(\frac{1}{2}-\frac{\beta}{2}+iq-ix\right)F_x(\varrho;x)-(1+\varrho)^{-a}\right)\right|\leq C_{q,\varrho_0, N} \varrho(1+\varrho)^{-\frac{3}{2}},

   with

   \left|\frac{1}{e^{\frac{\pi}{2}x}|\Gamma(\frac{1}{2}-\frac{\beta}{2}+iq-ix)|}+1\right|\leq C_{q,\beta,N} (x+1)^{-2}e^{-\frac{\pi}{2}x}.

2. Let \varrho\geq \varrho_0 and let \epsilon>0 be arbitrarily small. Then for \varrho_0 suitably small and for all N\in \mathbb{N}_0, there exists a constant C_{N,q,\varrho_0,\epsilon}>0, such that for all 0\leq m,l\leq N :

   \begin{aligned} \Bigg|(x\partial_{x})^l(\varrho\partial_{\varrho})^m&\left(\frac{\sin (\pi \textnormal{Im}\, \beta)}{\pi}F(\varrho;x)-\frac{\varrho^{-a}}{\Gamma(b)\Gamma(c_x-a)}+\frac{\varrho^{-b}}{\Gamma(a)\Gamma(c_x-b)}\right)\Bigg|\\ \leq &\: C_{N,q, \varrho_0,\epsilon} \varrho^{-1} e^{(\frac{\pi}{2}+\epsilon) x}\quad(\beta\in i(0,\infty)),\\ \Bigg|(x\partial_{x})^l(\varrho\partial_{\varrho})^m&\left(F(\varrho;x)-\frac{\varrho^{-a}}{\Gamma(a)\Gamma(c_x-a)}\left(\log \varrho+2\gamma_{\rm Euler}-\frac{\Gamma'(a)}{\Gamma(a)}-\frac{\Gamma'(c_x-a)}{\Gamma(c_x-a)}\right)\right)\Bigg|\\ \leq &\: C_{N,q, \varrho_0,\epsilon} \varrho^{-1}\log \varrho e^{(\frac{\pi}{2}+\epsilon) x}\quad (\beta=0). \end{aligned}

Let \beta_{\ell}\in i(0,\infty) and \mathfrak{w}_{\ell}^0(r)=\alpha_+ (\beta_{\ell})r^{\frac{1}{2}+\frac{\beta_{\ell}}{2}+iq}+\alpha_-(\beta_{\ell})r^{\frac{1}{2}-\frac{\beta_{\ell}}{2}+iq}, with \alpha_{\pm}(is)\in \mathbb{C} for s\in \mathbb{R} satisfying \alpha_-(is)=\alpha_+(-is).

Define

so that \zeta(-\beta_{\ell})=\frac{1}{\zeta(\beta_{\ell})}, and assume that there exists a \eta_0=\eta(|q|)>0, such that

Then the following sum is a well-defined smooth function with respect to (\tau,\varrho_{\infty},\theta,\varphi):

The function \Psi^{\infty}_0=(\Psi^{\infty}_0)_{\ell m}Y_{\ell m} is a solution to (?) with A=\widehat{A} and \mathbbm{h}=0.

Furthermore, f(\tau,r):=(\mathfrak{w}^0_{\ell})^{-1}(r)(\Psi^{\infty}_0)_{\ell m}(\tau,r) satisfies: for all k,l\leq N

Let \beta_{\ell}=0 and \alpha_+,\alpha_-\in \mathbb{C}. Let \mathfrak{w}_{\ell}^0(r)=r^{\frac{1}{2}+iq}(\alpha_+ \log r+\alpha_-).

Define

with

and with \gamma_{\rm Euler} the Euler–Mascheroni constant. Assume that there exists a \eta_0=\eta(|q|)>0, such that

Then \Psi^{\infty}_0=(\Psi^{\infty}_0)_{\ell m}Y_{\ell m} is well-defined and smooth with respect to the coordinate chart (\tau,\varrho_{\infty},\theta,\varphi) and it is a solution to (?) with A=\widehat{A} and \mathbbm{h}=0.

Furthermore, f(\tau,r):=(\mathfrak{w}^0_{\ell})^{-1}(r)(\Psi^{\infty}_0)_{\ell m}(\tau,r) satisfies: for all k,l\leq N

The functions (\Psi_0^{\infty})_{\ell m}(\tau,r), defined in Propositions A.2, A.4 and A.5 satisfy the following late-time asymptotics:

1. In the limit r\to \infty:

   \begin{aligned} (\Psi_0^{\infty})_{\ell m}(\tau,\infty)=&\:\beta_{\ell}(1+\tau)^{-\frac{1}{2}-\frac{\beta_{\ell}}{2}+iq}\sum_{n=0}^{N_{\beta_{\ell}}}\frac{\zeta^n(\beta_{\ell})(1+\tau)^{-n\beta_{\ell}}}{\Gamma\left(\frac{1}{2}+iq-(n+\frac{1}{2})\beta_{\ell}\right)}\\ =&\: \frac{\beta_{\ell}(1+\tau)^{-\frac{1}{2}-\frac{\beta_{\ell}}{2}+iq}}{\Gamma\left(\frac{1}{2}+iq-\frac{1}{2}\beta_{\ell}\right)}+O((1+\tau)^{-\frac{1}{2}-\frac{3\beta_{\ell}}{2}}),\quad (\beta_{\ell}\in (0,1)),\\ (\Psi_0^{\infty})_{\ell m}(\tau,\infty)=&\:\,\textnormal{sign}\,(q)\beta_{\ell}(1+\tau)^{-\frac{1}{2}-\,\textnormal{sign}\,(q)\frac{\beta_{\ell}}{2}+iq}\sum_{n=0}^{\infty}\frac{\zeta^n(\,\textnormal{sign}\,(q)\beta_{\ell})(1+\tau)^{-\,\textnormal{sign}\,(q)n\beta_{\ell}}}{\Gamma\left(\frac{1}{2}+iq-(n+\frac{1}{2})\,\textnormal{sign}\,(q)\beta_{\ell}\right)},\quad (\beta_{\ell}\in i(0,\infty)),\\ (\Psi_0^{\infty})_{\ell m}(\tau,\infty)=&\:\frac{-i}{\Gamma\left(\frac{1}{2}+iq\right)}(\tau+1)^{-\frac{1}{2}+iq}(\log(1+\tau))^{-1}+(\log(1+\tau))^{-2}O_{\infty}((1+\tau)^{-\frac{1}{2}}),\quad (\beta_{\ell}=0). \end{aligned}

2. At r=r_0>0:

   \begin{aligned} (\Psi^{\infty}_0)_{\ell m}(\tau,r_0)=&-2^{\frac{1}{2}+\frac{1}{2}\beta_{\ell}}\frac{\pi \beta_{\ell}}{\alpha_+\sin (\pi \beta_{\ell})}(1+\tau)^{-1-\beta_{\ell}}\mathfrak{w}_{\ell}^0(r)\\ \times & \left[\sum_{n=0}^{N_{\beta_{\ell}}}\zeta^n(\beta_{\ell}) \frac{(1+\tau)^{-n\beta_{\ell}}}{\Gamma(\beta_{\ell}+1) \Gamma(\frac{1}{2}-\frac{1}{2}\beta_{\ell}+iq)\Gamma(-(n+1)\beta_{\ell})}\right]\\ +&\:O_{\infty}\left((1+\tau)^{-2}\right)=-2^{\frac{1}{2}+\frac{1}{2}\beta_{\ell}}\frac{\pi \beta_{\ell}}{\alpha_+\sin (\pi \beta_{\ell}) \Gamma(\beta_{\ell}+1) \Gamma(\frac{1}{2}-\frac{1}{2}\beta_{\ell}+iq)\Gamma(-\beta_{\ell})}\mathfrak{w}_0(r)(1+\tau)^{-1-\beta_{\ell}}\\ &+O_{\infty}\left((1+\tau)^{-1-2\beta_{\ell}}\right),\quad (\beta_{\ell}\in (0,1)),\\ (\Psi_0^{\infty})_{\ell m}(\tau,\infty)=&\:-2^{\frac{1}{2}+\frac{1}{2}\,\textnormal{sign}\,(q)\beta_{\ell}}\frac{\pi \beta_{\ell}}{\alpha_-\sin (\pi \beta_{\ell})}(1+\tau)^{-1}\mathfrak{w}_{\ell}^0(r)\\ \times& \left[\sum_{n=0}^{\infty}\zeta^n(\,\textnormal{sign}\,(q)\beta_{\ell}) \frac{(1+\tau)^{-\,\textnormal{sign}\,(q)n\beta_{\ell}}}{\Gamma(\,\textnormal{sign}\,(q)\beta_{\ell}+1) \Gamma(\frac{1}{2}-\frac{1}{2}\,\textnormal{sign}\,(q)\beta_{\ell}+iq)\Gamma(-(n+1)\,\textnormal{sign}\,(q)\beta_{\ell})}\right]\\ +&\:O_{\infty}\left((1+\tau)^{-2}\right),\quad (\beta_{\ell}\in i(0,\infty)),\\ (\Psi^{\infty}_0)_{\ell m}(\tau,r_0)=&-\frac{2^{\frac{1}{2}+iq}}{\Gamma(\frac{1}{2}+iq)}\mathfrak{w}_{\ell}^0(r)\left[\log(1+\tau))^{-2}(\tau+1)^{-1-\,\textnormal{sign}\,(q)\beta_{\ell}}+(\log(1+\tau))^{-3}O_{\infty}((\tau+1)^{-1}\right]\\ +&\:O_{\infty}((1+\tau)^{-2}),\quad (\beta_{\ell}=0). \end{aligned}

Define the following constants: let k\in \mathbb{N}_0, then

1. The following bounds hold for (r^2X)^{k+1}(\Psi^{\infty}_0)_{\ell m} with k\in \mathbb{N}_0: there exist constants C=C(q,\beta_{\ell})>0 and \widetilde{C}=\widetilde{C}(q)>0, such that:

   \begin{aligned} \left|(r^2X)^{k+1}(\Psi^{\infty}_0)_{\ell m}(\tau,\infty)-\mathfrak{c}_k (1+\tau)^{\frac{k+1}{2}+iq-\frac{\beta_{\ell}}{2}}\right|\leq &\: C (1+\tau)^{\frac{k+1}{2}-\frac{3\beta_{\ell}}{2}}\quad (\beta_{\ell}\in (0,1)),\\ \limsup_{\tau\to \infty}(1+\tau)^{-\frac{k+1}{2}}|(r^2X)^{k+1}(\Psi^{\infty}_0)_{\ell m}|(\tau,\infty)\geq &\:\mathfrak{c}_k \quad (\beta_{\ell}\in i(0,\infty)),\\ |(r^2X)^{k+1}(\Psi^{\infty}_0)_{\ell m}|(\tau,\infty)\leq &\: C (1+\tau)^{\frac{k+1}{2}} \quad (\beta_{\ell}\in i(0,\infty)) ,\\ \left|(r^2X)^{k+1}(\Psi^{\infty}_0)_{\ell m}(\tau,\infty)-\mathfrak{c}_k \frac{(1+\tau)^{\frac{k+1}{2}+iq}}{\log(1+\tau)}\right|\leq &\: \widetilde{C}(\log(1+\tau))^{-2}(1+\tau)^{\frac{k+1}{2}}\quad (\beta_{\ell}=0). \end{aligned}

2. Let R>0 be arbitrary. Then there exist constants C=C(q,\beta_{\ell},R)>0 and \widetilde{C}=\widetilde{C}(q,R)>0, such that:

   \begin{aligned} \left|\int_0^{\infty}r^2|X(\Psi^{\infty}_0)_{\ell m}|^2(\tau,r)\,dr-\mathfrak{c}_{\rm int} (1+\tau)^{-\beta_{\ell}}\right|\leq &\: C (1+\tau)^{-2\beta_{\ell}}\quad (\beta_{\ell}\in (0,1)),\\ \limsup_{\tau\to \infty}\int_0^{\infty}r^2|X(\Psi^{\infty}_0)_{\ell m}|^2(\tau,r)\,dr\geq &\:\mathfrak{c}_{\rm int} \quad (\beta_{\ell}\in i(0,\infty)),\\ \sup_{\tau\in [0,\infty)}\int_0^{\infty}r^2|X(\Psi^{\infty}_0)_{\ell m}|^2(\tau,r)\,dr\leq &\: C \quad (\beta_{\ell}\in i(0,\infty)) ,\\ \left|\int_0^{\infty}r^2|X(\Psi^{\infty}_0)_{\ell m}|^2(\tau,r)\,dr-\frac{\mathfrak{c}_{\rm int}}{\log^2(1+\tau)}\right|\leq &\: \widetilde{C}(\log(1+\tau))^{-3}\quad (\beta_{\ell}=0). \end{aligned}

3. Let p\in (0,2] and consider the energy density \mathcal{E}_p from § ? and with \Omega^2=1-2Mr^{-1}+Q^2, Q^2\leq M^2. Let N\in \mathbb{N}_0. Then there exist constants C_N=C_N(q,\beta_{\ell},p,N)>0 and \widetilde{C}_N=\widetilde{C}_N(q,R,p,N)>0, such that for all k_1+k_2+k_3\leq N:

   \int_{r_+}^{\infty}\int_{\mathbb{S}^2}\mathcal{E}_{p}[\slashed{\nabla}_{\mathbb{S}^2}^{k_1}(rX)^{k_2}((\tau+1)T)^{k_3}(\Psi^{\infty}_0)]\,d\sigma dr\leq \begin{cases} C_{N}(\tau+1)^{p-2-\textnormal{Re}\, \beta_{\ell}}\quad &(\beta_{\ell}\neq 0),\\ \widetilde{C}_N\frac{(\tau+1)^{p-2}}{\log^2(1+\tau)}\quad &(\beta_{\ell}= 0). \end{cases}

Let \alpha_+,\alpha_-\in \mathbb{C} be defined as in Proposition 3.1. Then (?) and (?) are satisfied.