VSI electromagnetic fields
Extracted source text · arXiv 1605.03727
\documentclass{ws-procs975x65}
% optional packages
%\usepackage{graphicx}
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
% a few author defined macros like:
\def\beq{\begin{equation}}
\def\eeq{\end{equation}}
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
\begin{document}
\def \d {{\rm d}}
\def \bm #1 {\mbox{\boldmath{$m_{(#1)}$}}}
\def \bF {\mbox{\boldmath{$F$}}}
\def \bV {\mbox{\boldmath{$V$}}}
\def \bff {\mbox{\boldmath{$f$}}}
\def \bT {\mbox{\boldmath{$T$}}}
\def \bk {\mbox{\boldmath{$k$}}}
\def \bl {\mbox{\boldmath{$\ell$}}}
\def \bn {\mbox{\boldmath{$n$}}}
\def \bbm {\mbox{\boldmath{$m$}}}
\def \tbbm {\mbox{\boldmath{$\bar m$}}}
%\def \w1 {W^{(1)}}
%\def \w0 {\mbox{{$W^{(0)}$}}} %{{W^{(0)}}}
\def \0 {^{(0)}}
\def \1 {^{(1)}}
\def \a {\alpha}
\def \B {\beta}
\def \T {\bigtriangleup}
\newcommand{\msub}[2]{m^{(#1)}_{#2}}
\newcommand{\msup}[2]{m_{(#1)}^{#2}}
\newcommand{\be}{\begin{equation}}
\newcommand{\ee}{\end{equation}}
\newcommand{\beqn}{\begin{eqnarray}}
\newcommand{\eeqn}{\end{eqnarray}}
\newcommand{\AdS}{anti--de~Sitter }
\newcommand{\AAdS}{\mbox{(anti--)}de~Sitter }
\newcommand{\AAN}{\mbox{(anti--)}Nariai }
\newcommand{\AS}{Aichelburg-Sexl }
\newcommand{\pa}{\partial}
\newcommand{\pp}{{\it pp\,}-}
\newcommand{\ba}{\begin{array}}
\newcommand{\ea}{\end{array}}
\newcommand{\M}[3] {{\stackrel{#1}{M}}_{{#2}{#3}}}
%\newcommand{\m}[3] {\!\!{\stackrel{\hspace{.3cm}#1 \ 0}{m}}_{\!\!\!{#2}{#3}}\,}
%\newcommand{\m}[3] {\!\!{\stackrel{\hspace{.3cm}#1}{m}}_{\!\!\!{#2}{#3}}\,}
\newcommand{\m}[3] {{\stackrel{\hspace{.3cm}#1}{m}}_{\!{#2}{#3}}\,}
\newcommand{\tr}{\textcolor{red}}
\newcommand{\tb}{\textcolor{blue}}
\newcommand{\tg}{\textcolor{green}}
\newcommand{\tor}{\textcolor{orange}}
\def\a{\alpha}
\def\g{\gamma}
\def\de{\delta}
%\def\b{\tb{\beta}}
\def\b{{\kappa_0}}
%\def\E{{\cal E}}
%\def\B{{\cal B}}
\def\R{{\cal R}}
\def\F{{\cal F}}
\def\L{{\cal L}}
\def\e{e}
\def\bb{b}
\title{VSI electromagnetic fields}
\author{Marcello Ortaggio and Vojt\v ech Pravda}
\address{Institute of Mathematics of the Czech Academy of Sciences \\ \v Zitn\' a 25, 115 67 Prague 1, Czech Republic \\
$^*$E-mail: ortaggio(at)math(dot)cas(dot)cz, pravda@math.cas.cz}
\begin{abstract}
A $p$-form $\bF$ is VSI (i.e., all its scalar invariants of arbitrary order vanish) in a $n$-dimensional spacetime {\em if and only if} it is of type N, its multiple null direction $\bl$ is ``degenerate Kundt'', and {$\pounds_{\bl}\bF=0$}. This recent result is reviewed in the present contribution and its main consequences are summarized. In particular, a subset of VSI Maxwell fields possesses a {\em universal} property, i.e., they also solve (virtually) any generalized (non-linear and with higher derivatives) electrodynamics, possibly also coupled to Einstein's gravity.
\end{abstract}
\bodymatter
\section{Introduction}
\label{sec_intro}
The present contribution is summary of the main results of our recent work \cite{OrtPra15}.
It is useful to start by defining the VSI property for a general tensor, i.e.,
\begin{definition}[VSI tensors]
\label{def_VSI}
A tensor in an $n$-dimensional spacetime with metric $g_{ab}$ is VSI$_I$ if all the scalar polynomial invariants constructed from the tensor itself and its covariant derivatives up to order $I$ ($I=0,1,2,3,\dots$) vanish. It is VSI if all its scalar polynomial invariants of {\em arbitrary order} vanish.
\end{definition}
As a generalization of the notion of null fields \cite{syngespec} to arbitrary $p$-forms, it is natural to introduce the following
\begin{definition}[$p$-forms of type N]
\label{def_N}
At a spacetime point, a $p$-form $\bF$ is of type~N if it satisfies
\be
\ell^a F_{a b_1\ldots b_{p-1}}=0 , \qquad \ell_{[a}F_{b_1\ldots b_{p}]}=0 ,
\label{BelDeb}
\ee
where $\bl$ is a null vector (this follows from \eqref{BelDeb} and need not be assumed). The second condition can be equivalently replaced by $\ell^a\, {}^*F_{a b_1\ldots b_{n-p-1}}=0$.
\end{definition}
This is equivalent to the type N condition in the set-up of \cite{Milsonetal05}.
First, with the results of\cite{Hervik11}, it is easy to see that {\em a $p$-form $\bF$ is VSI$_0$ iff it is is of type N}. Next, our main result is the VSI condition, given in the next section.
\section{Main result: VSI $p$-forms}
The main result of\cite{OrtPra15} is the following
\begin{theorem}[VSI $p$-forms\cite{OrtPra15}]
\label{theor}
The following two conditions are equivalent:
\begin{enumerate}
\item\label{cond1} a non-zero $p$-form field $\bF$ is VSI in a spacetime with metric $g_{ab}$
\item\label{cond2}
\begin{enumerate}
\item\label{cond2a} $\bF$ possesses a multiple null direction $\bl$, i.e., it is of type N
\item\label{cond2b} {$\pounds_{\bl}\bF=0$} %$\nabla_{\bl}\bF=0$
\item\label{cond2c} $g_{ab}$ is a {\em degenerate Kundt} metric, and $\bl$ is the corresponding Kundt null direction.
\end{enumerate}
\end{enumerate}
\end{theorem}
We observe that, in Theorem~\ref{theor}, the $p$-form $\bF$ is not assumed to satisfy any particular field equations and the result is thus purely geometric (on the other hand, {\em if} $\bF$ is taken to be {\em closed}, i.e., $\d\bF=0$, then condition~\ref{cond2b} automatically follows from the type N condition \ref{cond2a}, and need not be assumed).
\subsection{Adapted coordinates}
From Theorem~\ref{theor} with the results of \cite{ColHerPel09a,Coleyetal09}, it follows that coordinates $(u,r,x^\alpha)$, adapted to $\bl=\partial_r$, exist such that any VSI $p$-form can be written as
\be
\bF=\frac{1}{(p-1)!}f_{\alpha_1\ldots\alpha_{p-1}}(u,x)\d u\wedge\d x^{\alpha_1}\wedge\ldots\wedge\d x^{\alpha_{p-1}} ,
\label{F_N_coords}
\ee
and the corresponding background metric as
\beqn
& & \d s^2 =2\d u\left[\d r+H(u,r,x)\d u+W_\alpha(u,r,x)\d x^\alpha\right]+ g_{\alpha\beta}(u,x) \d x^\alpha\d x^\beta , \label{Kundt_gen} \\
& & W_{\alpha}(u,r,x)=rW_{\alpha}^{(1)}(u,x)+W_{\alpha}^{(0)}(u,x) , \label{deg_Kundt1} \\
& & H(u,r,x)=r^2H^{(2)}(u,x)+rH^{(1)}(u,x)+H^{(0)}(u,x) , \label{deg_Kundt2}
\eeqn
where $\alpha,\beta=2 \dots n-1$, and $x$ denotes collectively the set of coordinates $x^\alpha$.
The degenerate Kundt metric~\eqref{Kundt_gen}--\eqref{deg_Kundt2} includes, e.g., all VSI spacetimes\cite{Pravdaetal02,Coleyetal04vsi}, all \pp waves, and all Kundt Einstein (or aligned pure radiation) spacetimes -- in particular, spacetimes of constant curvature (Minkowski and (A)dS). The corresponding Weyl and Ricci tensors are both of aligned type II, in general.
If $\bF$ in \eqref{F_N_coords} is required to obey the source-free Maxwell equations $\d\bF=0=\d^*\bF$, the following conditions are obtained
\be
f_{[\a_2\ldots\a_{p-1},\a_1]}=0 , \qquad (\sqrt{\tilde g}\,f^{\beta\a_1\ldots\a_{p-2}})_{,\beta}=0 ,
\label{Maxwell}
\ee
where $\tilde g\equiv\det g_{\alpha\beta}=-\det g_{ab}\equiv -g$. These are Maxwell's equations for the $(p-1)$-form $\bff$ in the $(n-2)$-dimensional Riemannian geometry associated with $g_{\alpha\beta}$, i.e., $\bff$ must be {\em harmonic} (recall, however, that $\bff$ can also depend on $u$). For $n=3,4$, it can be argued that to any $\bF$ which is VSI, it can always be associated a solution of the Maxwell equations which is also VSI\cite{OrtPra15}. It is also interesting to observe that the effective Maxwell equations \eqref{Maxwell} are ``immune'' to adding a Chern-Simons term (except when this is linear, see\cite{OrtPra15} for details) to the full Maxwell equations.
\subsection{Universal solutions of generalized electrodynamics}
\label{subsec_univ_test}
It was already known to Schr{\"o}dinger \cite{Schroedinger35,Schroedinger43} that all null Maxwell fields (with $n=4=2p$) solve the equations for the electromagnetic field in any non-linear electrodynamics. More generally, we observe that a subset of the VSI Maxwell fields described above possesses a ``universal'' property, i.e., they solve simultaneously any electrodynamics whose field equations can be expressed as $\d\bF=0$, $*\d\!*\!\!\tilde\bF=0$, where $\tilde\bF$ can be any $p$-form constructed from $\bF$ and its covariant derivatives. For example, any VSI Maxwell $\bF$ is universal if the background is a Kundt spacetime of Weyl and traceless-Ricci type III (aligned) with $DR=0=\delta_i R$ (an affine parameter and a frame parallelly transported along $\bl$ are assumed -- cf.\cite{OrtPra15} for the notation employed here). In particular, Ricci flat and Einstein Kundt spacetimes of Weyl type III/N/O can occur, the latter including Minkowski and (A)dS.
An explicit simple example is given by the Maxwell field
\be
\bF=e^{x/2}c(u)\d u\wedge\left(-\cos\frac{ye^u}{2}\d x+e^u\sin\frac{ye^u}{2}\d y\right) ,
\label{F_example}
\ee
defined in the type III vacuum spacetime found by Petrov (eq.~(31.40) in \cite{Stephanibook})
\be
\d s^2 =2\d u\left[\d r+\frac{1}{2}\left(xr-xe^x\right)\d u\right]+ e^x(\d x^2+e^{2u}\d y^2) . \label{Petrov} \\
\ee
More general results will be presented elsewhere.
\section{Einstein-Maxwell solutions}
\label{sec_EM}
\subsection{General field equations}
The previous discussion applies to VSI test fields, since we have not considered the consequences of the backreaction on the spacetime geometry. In the full Einstein-Maxwell theory with an arbitrary cosmological constant $\Lambda$, one finds that
the metric functions entering \eqref{Kundt_gen} must satisfy the following set of equations (obtained in\cite{OrtPra15} refining the results of\cite{PodZof09})
\beqn
& & \R_{\alpha\beta}=\frac{2\Lambda}{n-2}g_{\alpha\beta}+\frac{1}{2}W_{\alpha}^{(1)}W_{\beta}^{(1)}-W_{(\alpha||\beta)}^{(1)} , \label{Rij} \\
& & 2H^{(2)}=\frac{\R}{2}-\frac{n-4}{n-2}\Lambda+\frac{1}{4}W^{(1)\alpha}W_{\alpha}^{(1)} , \label{H2} \\
& & W_{\a||\B}^{(1)\ \B}=\frac{1}{2}W^{(1)\B}\left(3W_{\a||\B}^{(1)}-W_{\B||\a}^{(1)}\right)+W^{(1)}_{\a}\left(W^{(1)\B}_{\ \ \ \ \ ||\B}-\frac{1}{2}W^{(1)\B}W^{(1)}_\B -\frac{2 \Lambda}{n-2}\right) , \\
& & 2H^{(1)}_{,\a}=-{g_{\a\B,u}}^{||\B}+2W^{(0)\ \ \ \B}_{[\a||\B]}-2W^{(0)\B}W^{(1)}_{\a||\B}+(W^{(0)\B}W^{(1)}_\B)_{,\a}+W^{(1)}_{\a,u}+2(\ln\sqrt{\tilde g})_{,u\a} \nonumber \\
& & \quad {}+W^{(1)}_\a\left[W^{(0)\B}W^{(1)}_\B-W^{(0)\B}_{\ \ \ \ \ ||\B}+(\ln\sqrt{\tilde g})_{,u}\right]+\frac{4\Lambda}{n-2}W^{(0)}_\a , \label{H1} \\
& & \Delta H^{(0)}+W^{(1)\a}H^{(0)}_{,\a}+W^{(1)\a}_{\ \ \ \ \ ||\a}H^{(0)}=W^{(0)\B}W^{(0)}_\B\left(\frac{1}{2}W^{(1)\a}_{\ \ \ \ \ ||\a}-\frac{2\Lambda}{n-2}\right) \nonumber \\
& & \quad {}+H^{(1)}\left[W^{(0)\a}_{\ \ \ \ \ ||\a}-(\ln\sqrt{\tilde g})_{,u}\right]-\frac{1}{2}(W^{(0)\a}W^{(1)}_\a)^2+W^{(0)[\a||\B]}W^{(0)}_{[\a||\B]}+W^{(0)\ ||\a}_{\a,u} \nonumber \\
& & \quad {}-W^{(0)\B}\left(2W^{(1)\a}W^{(0)}_{[\a||\B]}+W^{(1)}_{\B,u}-2H^{(1)}_{,\B}\right)-(\ln\sqrt{\tilde g})_{,uu}+\frac{1}{4}g^{\a\B}_{\ \ ,u}g_{\a\B,u} -\b\F^2 . \label{H0}
\eeqn
Heree $\R_{\alpha\beta}$, $\R$ and $||$ denote, respectively, the Ricci tensor, the Ricci scalar and the covariant derivative associated with $g_{\alpha\beta}$, $W^{(1)\alpha}\equiv g^{\alpha\beta}W_{\beta}^{(1)}$, $\Delta$ is the Laplace operator in the geometry of the transverse metric $g_{\a\B}$, and $\b$ is a gravitational coupling constant. The Maxwell equations~\eqref{Maxwell} must also be satisfied.
The simplest examples one can construct are electromagnetic and gravitational ``plane-fronted'' waves (with $W_{\alpha}^{(0)}=0$) propagating in a constant curvature background, giving rise to Kundt waves of Weyl type N (in four dimensions see, e.g., \cite{GarPle81,OzsRobRoz85}). More general (e.g., with $W_{\alpha}^{(0)}\neq0$) degenerate Kundt metrics with null Maxwell fields are also known (see \cite{Stephanibook,GriPodbook,GriDocPod04} and references therein for $n=4$). The case of VSI $p$-form Maxwell fields in VSI and \pp waves spacetimes have been discussed in\cite{OrtPra15} (where further references can be found).
\subsection{Universal Einstein-Maxwell solutions}
\label{subsec_univ_EM}
Some of the universal Maxwell fields mentioned in section~\ref{subsec_univ_test} can also be used to construct exact solutions of full general relativity, where the energy-momentum tensor $T_{ab}$ associated with the electromagnetic field is determined in the generalized electrodynamics (in terms of $\bF$ and its covariant derivatives -- cf. \cite{Podolsky42} for an example). For example, as pointed out in\cite{OrtPra15}, all VSI spacetimes with $L_{i1}=0=L_{1i}$ (i.e., the recurrent ones) coupled to an aligned VSI $p$-form field that solve the standard Einstein-Maxwell equations are also exact solutions of gravity coupled to generalized electrodynamics, provided $p>1$ and $\delta_iF_{1j_1\ldots j_{p-1}}=0$ (in an ``adapted'' parallely transported frame, i.e., such that $\M{i}{j}{k}=0$). Within this family, metrics of Weyl type N are necessarily \pp waves, for which such a universal property {was pointed out in} \cite{Guven87,HorSte90,Horowitz90}, at least for certain values of $p$. But metrics of Weyl type III are also permitted, including \pp waves ($L_{11}=0$) and also genuinely recurrent ($L_{11}\neq0$) spacetimes (for $n=4$, $p=3$ this was discussed in \cite{Coley02}). One explicit example of the latter solutions in 4D is given by the Maxwell field~\eqref{F_example} with the metric
\be
\d s^2 =2\d u\left[\d r+\frac{1}{2}\left(xr-xe^x-2\b e^xc^2(u)\right)\d u\right]+ e^x(\d x^2+e^{2u}\d y^2) , \label{Petrov_EM}
\ee
which is a modification of \eqref{Petrov} taking into account the backreaction.
As in section~\ref{subsec_univ_test}, the above discussion applies to generalized electrodynamics with arbitrary higher-order derivative ``corrections''. A special instance of this result is the fact that Einstein-Maxwell solutions with aligned null electromagnetic fields (not necessarily VSI) are also solution of NLE coupled to gravity, as previously demonstrated in \cite{Kichenassamy59,KreKic60,Peres61}.
\section{Further remarks}
For certain purposes, it may be useful to observe that if $\bF$ is VSI$_3$ then it is necessarily VSI (this follows from the proof of Theorem~\ref{theor} given in\cite{OrtPra15}). For completeness, let us thus also give the necessary and sufficient conditions for a $p$-form $\bF$ to be VSI$_1$ or VSI$_2$ (recall that VSI$_0$ means type N, as mentioned in section~\ref{sec_intro}):
\begin{proposition}[VSI$_1$ and VSI$_2$ $p$-forms\cite{OrtPra15}]
\label{proposition_VSI1_2}
A $p$-form $\bF$ is VSI$_1$ iff it is is of type N, {$\pounds_{\bl}\bF=0$}, $\bl$ is Kundt. It is VSI$_2$ iff it is is of type N, {$\pounds_{\bl}\bF=0$}, $\bl$ is Kundt and (at least) doubly aligned with the Riemann tensor.
\end{proposition}
It is also interesting to observe that Theorem~\ref{theor} and Proposition~\ref{proposition_VSI1_2} apply also in the limiting case $p=1$, i.e., when $\bF$ is a vector field. In particular, when $p=1$, Theorem~\ref{theor} reduces to: {\em a vector field $\bl$ is VSI in a spacetime with metric $g_{ab}$ iff $\bl$ is Kundt and affinely parameterized, and $g_{ab}$ is a degenerate Kundt metric w.r.t $\bl$}.
The above comments are summarized in the first two columns of Fig.~\ref{table}. The last column gives corresponding results in the case of the Riemann tensors (i.e., for the VSI$_I$ spacetimes)\cite{Pravdaetal02,Coleyetal04vsi,Pelavasetal05}. While certain conditions turn out to be similar, there is also an important difference: for the Riemann tensor one has already $VSI_2\Rightarrow VSI$.
\begin{figure}[t]
%\begin{table}[t]
\begin{center}
\begin{tabular}{|c|l|l|l|}
\hline
&&&\\[-3mm]
& \ vector $\bl$ \ & \ \ $p$-form $\bF$ \ \ & \ \ Riemann \ \ \\
&&&\\[-3mm]
\hline &&&\\[-10pt]
VSI$_0$ & \hspace{-.25cm} null & \hspace{-.25cm} N & \hspace{-.25cm} III (N,O) \\[1pt]
\hline &&&\\[-10pt]
VSI$_1$ & \hspace{-.25cm} Kundt & \hspace{-.25cm} N, $\pounds_{\bl}\bF=0$, Kundt & \hspace{-.25cm} (*) N, \!$\kappa=0$, $\sigma\Psi_4\!=\!\rho\Phi_{22}$ \\[1pt]
\hline &&&\\[-10pt]
VSI$_2$ & \hspace{-.25cm} Kundt, Riem II & \hspace{-.25cm} N, $\pounds_{\bl}\bF=0$, Kundt, Riem II & \hspace{-.25cm} III, Kundt \\[1pt]
\hline &&&\\[-10pt]
VSI$_3$ & \hspace{-.25cm} degKundt & \hspace{-.25cm} N, $\pounds_{\bl}\bF=0$, degKundt & \hspace{-.25cm} `` \\[1pt]
\hline &&&\\[-10pt]
\vdots & \hspace{-.25cm} `` \ & \hspace{-.25cm} `` \ & \hspace{-.25cm} `` \\[1pt]
\hline &&&\\[-10pt]
VSI & \hspace{-.25cm} `` \ & \hspace{-.25cm} `` \ & \hspace{-.25cm} `` \ \\[1pt]
\hline
\end{tabular}
\caption{Comparing various VSI$_I$ conditions for a 1-form $\bl$, a $p$-form $\bF$\cite{OrtPra15} and the Riemann tensor\cite{Pravdaetal02,Coleyetal04vsi,Pelavasetal05}. The quotation marks `` mean that the same conditions as in the next-upper row apply. The Riemann VSI$_1$ condition (*) needs some comments for two reasons. First, it has been investigated only in the case $n=4$\cite{Pelavasetal05} (and indeed it is given above in the standard 4D NP notation). Additionally, it is sufficient but not necessary (contrary to the rest of the table): also Kundt spacetimes of Riemann type III are VSI$_1$ (but not ``properly'', i.e., they are in fact VSI, cf. the lower rows in the same column).}
\label{table}
\end{center}
%\end{table}
\end{figure}
%----------------------------------------------------------------------------------------
%----------------------------------------------------------------------------------------
\section*{Acknowledgments}
We thank research plan {RVO: 67985840} and research grant GA\v CR 13-10042S.
\begin{thebibliography}{10}
\bibitem{OrtPra15}
M.~Ortaggio and V.~Pravda, Electromagnetic fields with vanishing scalar invariants, {\em Class. Quantum Grav.} {\bf 33}, p. 115010 (2016).
\bibitem{syngespec}
J.~L. Synge, {\em Relativity: the Special Theory} (North-Holland, Amsterdam,
1955).
\bibitem{Milsonetal05}
R.~Milson, A.~Coley, V.~Pravda and A.~Pravdov\'a, Alignment and algebraically
special tensors in {L}orentzian geometry, {\em Int. J. Geom. Meth. Mod.
Phys.} {\bf 2}, 41 (2005).
\bibitem{Hervik11}
S.~Hervik, A spacetime not characterized by its invariants is of aligned type
{II}, {\em Class. Quantum Grav.} {\bf 28}, p. 215009 (2011).
\bibitem{ColHerPel09a}
A.~Coley, S.~Hervik and N.~Pelavas, Spacetimes characterized by their scalar
curvature invariants, {\em Class. Quantum Grav.} {\bf 26}, p. 025013 (2009).
\bibitem{Coleyetal09}
A.~Coley, S.~Hervik, G.~O. Papadopoulos and N.~Pelavas, Kundt spacetimes, {\em
Class. Quantum Grav.} {\bf 26}, p. 105016 (2009).
\bibitem{Pravdaetal02}
V.~Pravda, A.~Pravdov\'a, A.~Coley and R.~Milson, All spacetimes with vanishing
curvature invariants, {\em Class. Quantum Grav.} {\bf 19}, 6213 (2002).
\bibitem{Coleyetal04vsi}
A.~Coley, R.~Milson, V.~Pravda and A.~Pravdov\'a, Vanishing scalar invariant
spacetimes in higher dimensions, {\em Class. Quantum Grav.} {\bf 21}, 5519
(2004).
\bibitem{Schroedinger35}
E.~Schr{\"{o}}dinger, Contributions to {B}orn's new theory of the
electromagnetic field, {\em Proc. Roy. Soc. London Ser. A} {\bf 150}, 465
(1935).
\bibitem{Schroedinger43}
E.~Schr{\"{o}}dinger, A new exact solution in non-linear optics
(two-wave-system), {\em Proc. Roy. Irish Acad.} {\bf A49}, 59 (1943).
\bibitem{Stephanibook}
H.~Stephani, D.~Kramer, M.~MacCallum, C.~Hoenselaers and E.~Herlt, {\em Exact
Solutions of {E}instein's Field Equations}, second edn. (Cambridge University
Press, Cambridge, 2003).
\bibitem{PodZof09}
J.~Podolsk\'y and M.~\v{Z}ofka, General {K}undt spacetimes in higher
dimensions, {\em Class. Quantum Grav.} {\bf 26}, p. 105008 (2009).
\bibitem{GarPle81}
A.~Garc\'{\i}a~D\'{\i}az and J.~F. Pleba\'nski, All nontwisting {N}'s with
cosmological constant, {\em J. Math. Phys.} {\bf 22}, 2655 (1981).
\bibitem{OzsRobRoz85}
I.~Ozsv\'ath, I.~Robinson and K.~R\'ozga, Plane-fronted gravitational and
electromagnetic waves in spaces with cosmological constant, {\em J. Math.
Phys.} {\bf 26}, 1755 (1985).
\bibitem{GriPodbook}
J.~B. Griffiths and J.~Podolsk\'y, {\em Exact Space-Times in {E}instein's
General Relativity} (Cambridge University Press, Cambridge, 2009).
\bibitem{GriDocPod04}
J.~Griffiths, P.~Docherty and J.~Podolsk\'y, Generalized {K}undt waves and
their physical interpretation, {\em Class. Quantum Grav.} {\bf 21}, 207
(2004).
\bibitem{Podolsky42}
B.~Podolsky, A generalized electrodynamics. {P}art {I} -- {N}on-quantum, {\em
Phys. Rev.} {\bf 62}, 68 (1942).
\bibitem{Guven87}
R.~G{\"u}ven, Plane waves in effective theories of superstrings, {\em Phys.
Lett. {\rm B}} {\bf 191}, 275 (1987).
\bibitem{HorSte90}
G.~T. Horowitz and A.~R. Steif, Spacetime singularities in string theory, {\em
Phys. Rev. Lett.} {\bf 64}, 260 (1990).
\bibitem{Horowitz90}
G.~T. Horowitz, Singularities in string theory, in {\em Strings '90\/}, eds.
R.~Arnowitt, R.~Bryan, M.~J. Duff, D.~Nanopoulos, C.~N. Pope and E.~Sezgin
(World Scientific, Singapore, 1990) pp. 163--175.
\bibitem{Coley02}
A.~A. Coley, A class of exact classical solutions to string theory, {\em Phys.
Rev. Lett.} {\bf 89}, p. 281601 (2002).
\bibitem{Kichenassamy59}
S.~Kichenassamy, Sur le champ \'electromagn\'etique singulier en th\'eorie de
{B}orn–-{I}nfeld, {\em C. R. Hebd. Seanc. Acad. Sci.} {\bf 248}, 3690–
(1959).
\bibitem{KreKic60}
H.~Kremer and S.~Kichenassamy, Sur le champ \'electromagn\'etique singulier
dans une th\'eorie du type {B}orn–-{I}nfeld, {\em C. R. Hebd. Seanc. Acad.
Sci.} {\bf 250}, 1192 (1960).
\bibitem{Peres61}
A.~Peres, Nonlinear electrodynamics in general relativity, {\em Phys. Rev.}
{\bf 122}, 273 (1961).
\bibitem{Pelavasetal05}
N.~Pelavas, A.~Coley, R.~Milson, V.~Pravda and A.~Pravdov\'a, {VSI}$_i$
spacetimes and the $\epsilon$-property, {\em J. Math. Phys.} {\bf 46}, p.
063501 (2005).
\end{thebibliography}
\end{document}