1: 

2: %%NEWEST VERSION--6 april 06

3: 

4: \documentclass{amsart}

5: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

6: \usepackage{graphicx,amssymb}

7: \newtheorem{thm}{Theorem}[section]

8: \newtheorem{cor}[thm]{Corollary}

9: \newtheorem{lem}[thm]{Lemma}

10: \newtheorem{prop}[thm]{Proposition}

11: \newtheorem{con}[thm]{Conjecture}

12: \theoremstyle{definition}

13: \newtheorem{defn}[thm]{Definition}

14: \theoremstyle{remark}

15: \newtheorem{rem}[thm]{Remark}

16: \def\beq{\begin{eqnarray}}

17: \def\eeq{\end{eqnarray}}

18: \def\bsp{\begin{split}}

19: \def\esp{\end{split}}

20: \def\lra{\longrightarrow}

21: \def\ra{\rightarrow}

22: \def\inl{\left\langle}

23: \def\inr{\right\rangle}

24: \def\dt{\partial_t}

25: \def\Tr{\mathrm{Tr}}

26: \def\d{\mathrm{d}}

27: \def\diag{\mathrm{diag}}

28: \def\RC{CSI_R}

29: \def\KC{CSI_K}

30: \def\FC{CSI_F}

31: \newcommand{\FCSI}[1]{CSI_{F,#1}}

32: \newcommand{\mf}[1]{{\mathfrak #1}}

33: \newcommand{\ul}[1]{\underline{#1}}

34: \newcommand{\mb}[1]{{\mathbb #1}}

35: \newcommand{\mc}[1]{{\cal #1}}

36: \newcommand{\mbold}[1]{\mbox{\boldmath{\ensuremath{#1}}}}

37: 

38: 

39: \begin{document}

40: 

41: \title[\textbf{CSI: Halifax}]{\textbf{On Spacetimes with Constant Scalar Invariants}}

42: \author{\textbf{Alan Coley, Sigbj\o rn Hervik, Nicos Pelavas}}

43: %EndAName

44: \address{Department of Mathematics and Statistics,

45: Dalhousie University, Halifax, Nova Scotia, Canada B3H 3J5}

46: \email{aac@mathstat.dal.ca, herviks@mathstat.dal.ca,

47: pelavas@mathstat.dal.ca}

48: \date{\today}

49: \maketitle

50: 

51: \begin{abstract}

52: 

53: We study Lorentzian spacetimes for which all scalar invariants

54: constructed from the Riemann tensor and its covariant derivatives

55: are constant ($CSI$ spacetimes). We obtain a number of general 

56: results in arbitrary dimensions. We study and construct

57: warped product $CSI$ spacetimes and higher-dimensional  Kundt

58: $CSI$ spacetimes. We show how these spacetimes can

59: be constructed from locally homogeneous spaces and $VSI$

60: spacetimes. The results suggest a number of conjectures. In

61: particular, it is plausible that for $CSI$ spacetimes that are not

62: locally homogeneous the Weyl type is $II$, $III$, $N$ or $O$, with any

63: boost weight zero components being constant.

64: We then consider the four-dimensional $CSI$ spacetimes in more

65: detail. We show that there are severe constraints on these

66: spacetimes, and we argue that it is plausible that they are either

67: locally homogeneous or that  the spacetime necessarily belongs to

68: the Kundt class of $CSI$ spacetimes, all of which are constructed.

69: The four-dimensional results lend support to the conjectures in

70: higher dimensions.

71: 

72: \end{abstract}

73: 

74: \newpage

75: \section{Introduction}

76: 

77: Let $(\mathcal{M},g)$ denote a differentiable manifold, either of

78: Riemannian or Lorentzian signature. We

79: are interested in all spacetimes for which all scalar invariants

80: constructed from the Riemann tensor and its covariant derivatives

81: are constant (denoted as $CSI$ spacetimes). In the case of a

82: Riemannian manifold, if it is $CSI$ then it is locally homogeneous

83: \cite{PTV}. This theorem is false for Lorentzian manifolds (there

84: are a number of counterexamples; e.g., the non-homogeneous $VSI$

85: spacetimes \cite{4DVSI}).

86: 

87: 

88: 

89: There are a number of $CSI$ spacetimes that arise from homogeneous

90: spacetimes and simple warped products (see section 3).  Some examples of $CSI$

91: spacetimes that arise as simple {\it tensor sum} generalizations of

92: $VSI$ spacetimes with a cosmological constant are given in

93: \cite{BijPod} (some higher dimensional versions are given in

94: \cite{obuk}). Simple examples of {\it fibered products} were given

95: in \cite{Siklos} and \cite{HJS} (see also \cite{Hervik}). It is also straightforward to

96: construct $CSI$ spacetimes as {\it warped products}. All of these

97: spacetimes belong to the class $\RC$ described

98: below.

99: 

100: 

101: We begin with some definitions, and a brief discussion of the

102: four-dimensional case.

103: 

104: 

105: 

106: \subsection{Notation}

107: 

108: Let us denote by $\mathcal{I}_k$ the set of all scalar

109: invariants constructed from the curvature tensor and its covariant derivatives up

110: to order $k$.

111: \begin{defn}[{$VSI$}$_k$ spacetimes]

112: $\mathcal{M}$ is called {$VSI$}$_k$ if for any invariant $I\in\mathcal{I}_k$, $I=0$ over $\mathcal{M}$.

113: \end{defn}

114: \begin{defn}[{$CSI$}$_k$ spacetimes]

115: $\mathcal{M}$ is called {$CSI$}$_k$ if for any invariant $I\in\mathcal{I}_k$, $\partial_{\mu}{I}=0$ over $\mathcal{M}$. 

116: \end{defn}

117: Moreover, if a spacetime is {$VSI$}$_k$ or {$CSI$}$_k$ for all

118: $k$, we will simply call the spacetime {$VSI$} or {$CSI$},

119: respectively. The set of all locally homogeneous spacetimes will be

120: denoted by $H$. Clearly $VSI \subset CSI$ and $H \subset CSI$.

121: 

122: \begin{defn}[{\sc $\RC$} spacetimes]

123: Let us denote by $\RC$

124: all reducible $CSI$ spacetimes that can be built from $VSI$ and

125: $H$ by (i) warped products (ii) fibered products, and (iii) tensor sums

126: (defined more precisely later).

127: \end{defn}

128: 

129: \begin{defn}[{\sc $\FC$} spacetimes]

130: Let us denote by $\FC$

131: those spacetimes for which there exists a frame with a null vector

132: $\ell$ such that all components of the Riemann tensor and its

133: covariants derivatives in this frame have the property that (i)

134: all positive boost weight components (with respect to $\ell$) are zero and (ii) all

135: zero boost weight components are constant.

136: \end{defn}

137: Note that $\RC \subset CSI$ and $\FC

138: \subset CSI$. (There are similar definitions for $\FCSI{k}$

139: etc. \cite{epsilon}).

140: 

141: \begin{defn}[{$\KC$} spacetimes]

142: Finally, let us denote by $\KC$, those $CSI$

143: spacetimes that belong to the (higher-dimensional) Kundt class (defined later);

144: the so-called Kundt $CSI$ spacetimes.

145: \end{defn}

146: 

147: In particular, we shall study the relationship between $\RC$,

148: $\FC$, $\KC$ and especially with

149: $CSI \backslash H$.  We note that by construction $\RC$ is at

150: least of Weyl type $II$ (i.e., of type $II$, $III$, $N$ or $O$

151: \cite{class}), and by definition $\FC$ and

152: $\KC$ are at least of Weyl type $II$ (more precisely,

153: at least of Riemann type $II$). In 4D,

154: $\RC$, $\FC$ and $\KC$ are

155: closely related, and it is plausible that $CSI \backslash H$ is at least of

156: Weyl type $II$ (see section 7).

157: 

158: In four dimensions the Weyl classification and the Petrov classification are closely related \cite{class}. However, due to the fact that the Weyl classification contains many more possibilities in dimensions higher than four we will  \emph{restrict the term Petrov classification to four dimensions only}.

159: 

160: \subsection{4D $CSI$}

161: 

162: We are particularly interested in the four-dimensional $CSI$

163: spacetimes. For a Riemannian manifold every $CSI$ is

164: homogeneous $(CSI \equiv H)$. This is not true for Lorentzian

165: manifolds.  However, for every $CSI$ with particular constant

166: invariants there is a homogeneous spacetime (not necessarily

167: unique) with precisely the same constant invariants.  This

168: suggests that $CSI$ can be ``built'' from $H$ and $VSI$ (e.g.,

169: $\RC$).

170: 

171: 

172: In addition, there is a relationship between the $CSI$ conditions

173: and (i) the rank of the Riemann tensor and its holonomy class  \cite{Hall} (ii)

174: the existence of curvature collineations \cite{Hall} (iii) the

175: condition of non-complete backsolvability\footnote{Complete baksolvability (CB) refers to when all of the components of the Riemann tensor can be determined from the CZ set \cite{Carminati} of zeroth order curvature invariants, once all the remaining frame freedom has been used to fix components of the Riemann tensor.} (NCB) \cite{Carminati}, in

176: addition to (iv) curvature homogeneity\footnote{$\mathcal{M}$ is curvature homogeneous (of order zero) if there exists a frame with respect to which the Riemann tensor has constant components.} and (v) sectional

177: curvature \cite{Hall}.

178: The relationship between $CSI$ and curvature homogeneity in $4D$

179: is studied in \cite{Mp}. In locally homogeneous spacetimes all of the sectional

180: curvatures (Gaussian curvatures) are constant. With the exception

181: of certain special plane wave and constant curvature spacetimes

182: (and all vacuum spacetimes), the sectional curvature uniquely

183: determines the spacetime metric \cite{Hall}.

184: 

185: It is clear that $CSI$ consists of $\RC$ and,

186: possibly, some other very special spacetimes. Let us present a

187: {\em heuristic argument} for this in $4D$. Let us suppose that the

188: spacetime is $CSI$.  For spacetimes that are completely

189: backsolvable (CB) there is a special invariant frame such that all

190: components of the Riemann tensor are constant. With respect to

191: this invariant frame there exist smooth vector fields $\zeta_i (i

192: = 1, 2, 3, 4)$ that act transitively on the manifold whose

193: directional derivatives leave the Riemann tensor invariant (i.e.,

194: $\zeta_i$ form a Lie algebra of curvature  collineations that span

195: $\mathcal{M}$).  In general, every curvature collineation is a homothetic

196: vector, so as a result there is a homothetic group acting

197: transitively on $\mathcal{M}$. Therefore, this then implies that in general

198: the $CSI$ spacetime is locally homogeneous.  The exceptions are

199: those spacetimes that are not CB (NCB \cite{Carminati}) and those

200: spacetimes for which a curvature collineation is not a homothetic

201: vector \cite{Hall}, and these spacetimes are related to the $VSI$

202: spacetimes.

203: 

204: 

205: \subsection{Overview}

206: 

207: In this paper we shall study $CSI$ spacetimes. In the

208: next section we shall begin by summarizing some important results.

209: In section 3 we construct a subclass of $\RC$

210: spacetimes that arise as warped products of a homogeneous space and a

211: $VSI$ spacetime. In section 4 we consider the higher-dimensional

212: Kundt class. In section 5 examples of $\KC$ spacetimes in

213: higher dimensions are given. We discuss the results in section 6,

214: and present a number of conjectures. In section 7 we summarize the

215: results in 4D in detail.

216: 

217: 

218: There are three appendices. In Appendix A, we present all of the

219: 3-dimensional $VSI$ metrics, which is necessary for the explicit

220: determination of the metrics in the 4-dimensional $\RC$ spacetimes. In

221: Appendix B,  4-dimensional $CSI$ spacetimes are given explicitly,

222: and the relationship between $\RC$,

223: $\FC$, and $\KC$ in 4D is discussed.

224: Finally, in Appendix C, we write down the metric for higher

225: dimensional $VSI$ spacetimes in a canonical form.

226: 

227: 

228: \section{Spacetimes with constant scalar invariants}

229: 

230: \subsection{Riemannian case}

231: Let us first consider the Riemannian case where a great deal about these spacetimes is known. The reason the Riemannian case is easier to deal with is because the orthogonal group $O(d)$ is compact, and hence, the orbits of its group action are also compact.

232: 

233: \begin{thm}

234: If $\mathcal{M}$ is a Riemannian spacetime, then:

235: \begin{enumerate}

236: \item{} {$VSI$}$_0$ $\Rightarrow$ {$VSI$}.

237: \item{} There exists a $k\in\mathbb{N}$ such that {$CSI$}$_k$ $\Rightarrow$ {$CSI$}.

238: \end{enumerate}

239: \end{thm}

240: \begin{proof}

241: (1): See proof below. \quad (2): See \cite{PTV}

242: \end{proof}

243: \begin{thm}[Riemannian {$VSI$}]

244: A Riemannian space is {$VSI$} if and only if it is flat.

245: \end{thm}

246: \begin{proof}

247: Consider the invariant $R_{ABCD}R^{ABCD}$. Using an orthonormal frame, this invariant is a sum of squares; hence, $R_{ABCD}R^{ABCD}=0$ $\Leftrightarrow$ $R_{ABCD}=0$.

248: \end{proof}

249: \begin{thm}[Riemannian {$CSI$}]

250: A Riemannian space is {$CSI$} if and only if it is locally homogeneous. Moreover, the set of curvature invariants, $\mathcal{I}$, uniquely determines the metric up to  isometries.

251: \end{thm}

252: \begin{proof}

253: See Pr\"ufer-Tricerri-Vanhecke \cite{PTV}.

254: \end{proof}

255: 

256: 

257: \begin{thm}

258: [Singer]  If $\mathcal{M}$ is curvature homogeneous of order $k \geq

259: \frac{1}{2} n(n-1)$, then $\mathcal{M}$ is locally homogeneous.

260: \end{thm}

261: \begin{proof}

262: See Singer \cite{Singer}.

263: \end{proof}

264: 

265: 

266: 

267: 

268: \subsection{Lorentzian case}

269: The Lorentzian case is more difficult to deal with and only

270: partial results are known. However, in the {$VSI$} case, it was

271: shown in \cite{Higher} that {$VSI$}$_2$ implies {$VSI$}. A

272: similar result in the {$CSI$} case is not known;

273: however, in this case we believe that an $n$ exists such that

274: {$CSI$}$_n$ implies {$CSI$}.

275: 

276: In the Riemannian case $\mathcal{M}$ is locally homogeneous if and only if $\mathcal{M}$ is

277: curvature homogeneous to all orders (Theorem 2.4). In the

278: Lorentzian case, we have that

279: \begin{thm}[Podesta and Spiro]

280: There is an integer $K_{p,q}$ such that if $(M,g)$ is a complete,

281: simply connected pseudo Riemannian manifold of type $(p,q)$ which

282: is $K_{p,q}$-curvature homogeneous, then $\mathcal{M}$ is locally

283: homogeneous.

284: \end{thm}

285: \begin{proof}

286: See Podesta and Spiro \cite{PS}.

287: \end{proof}

288: 

289: We argued above that in 4D Lorentzian manifolds for every $CSI$ with particular constant

290: invariants there is a locally homogeneous spacetime with 

291: the same constant invariants.  It is plausible that this is

292: true in higher dimensions.

293: 

294: 

295: \begin{con}

296: Assume that a  Lorentzian spacetime $\mathcal{M}$ is a {$CSI$} spacetime with

297: curvature invariants $\mathcal{I}$. Then there exists a locally homogeneous

298: space $\widetilde{\mathcal{M}}$ with curvature invariants

299: $\widetilde{\mathcal{I}}=\mathcal{I}$. 

300: \end{con}

301: 

302: 

303: 

304: 

305: 

306: \subsection{Boost-weight decomposition}

307: Consider an arbitrary covariant tensor $T$ and a null frame $\left\{\ell_{A},n_{A}, m^{\hat\mu}_{A}\right\}$; i.e.,

308: \[ \ell_{A}n^{A}=1,~~m^{\hat\mu}_{~A}m^{\hat\nu A}=\delta^{\hat\mu\hat\nu}, ~~ \ell_{A}\ell^{A}=n_{A}n^{A}=\ell_{A}m^{\hat\mu A}=

309: n_{A}m^{\hat\mu A}=0.\]

310: Consider now a boost in the $\ell^{A}-n^{A}$-plane:

311: \beq

312: \left\{\tilde{\ell}_{A},\tilde{n}_{A}, \tilde{m}^{\hat\mu}_{~A}\right\}= \left\{e^{\lambda}\ell_{A},e^{-\lambda}n_{A}, m^{\hat\mu}_{~A}\right\}.

313: \eeq

314: We can consider the vector-space decomposition of the tensor $T$ in terms of the boost weight with respect to the above boost (see \cite{Milson}):

315: \beq

316: T=\sum_b(T)_b,

317: \eeq

318: where $(T)_b$ denotes the projection of the tensor $T$ onto the vector space of boost-weight $b$. The components of the tensor $(T)_b$ with respect to the frame will transform according to:

319: \beq

320: (T)_{b~AB...}=e^{-b\lambda}(T)_{b~\tilde{A}\tilde{B}...}.

321: \eeq

322: Furthermore, we note that

323: \beq

324: \left(T\otimes S\right)_b=\sum_{b'+b''=b}(T)_{b'}\otimes(S)_{b''}.

325: \eeq

326: Moreover, the connection  $\nabla$ can similarly be decomposed according to whether it raises, lowers or preserves the boost weight:

327: \beq

328: \nabla=(\nabla)_{-1}+(\nabla)_0+(\nabla)_1,

329: \eeq

330: where $(\nabla)_b$ is defined by

331: \[ (\nabla)_{b{\bf X}}\equiv \nabla_{({\bf X})_b},\]

332: for all vectors ${\bf X}$. 

333: We note that for the metric, $g=(g)_0$; hence, raising or lowering tensor indices preserves the boost weight.

334: 

335: An invariant of a tensor, $T$, is necessarily $SO(1,n)$-invariant; in particular, for any full contraction of $T$:

336: \beq

337: \mathrm{Cont}[T]=\mathrm{Cont}[(T)_0].

338: \eeq

339: This property can now be utilized to construct spaces with constant curvature invariants. In the {$VSI$} case, there exists a null-frame such that the Riemann tensor, $R$, has the property:

340: \beq

341: (R)_b=0,\quad b\geq 0.

342: \eeq

343: This property alone implies that $\mathcal{I}_0=0$, regardless of $(R)_b$ ($b<0$). 

344: If the curvature tensor is of the form:

345: \beq

346: (R)_b=0,\quad b> 0, \quad \text{and}~(R)_{b~ABCD,E}=0,\quad b= 0,

347: \eeq

348: in a frame for which $g_{AB,E}=0$, it is necessarily a {$CSI$}$_0$ spacetime. In general, further restrictions must be put on

349: $\nabla$ in order to make it {$CSI$}. We note that spacetimes can can be classified according 

350: to their boost-weight components $(R)_b$. In particular, a classification using the Weyl 

351: tensor $C$, which generalises the Petrov classification in 4D, has been employed based on 

352: the components $(C)_b$ \cite{class,Milson}.

353: 

354: \subsection{Kundt metrics}

355: \label{sect:Kundt} In the 4D $VSI$ spacetimes, all of the positive

356: boost weight spin coefficients (e.g., $\kappa, \rho$ and $\sigma$)

357: are zero, and hence it follows that $\ell$ is geodesic,

358: non-expanding, shear-free and non-twisting and the spacetime is

359: Kundt $K$ \cite{exsol}. This is also true for $CSI$ spacetimes,

360: in the sense that if the $CSI$ spacetime is not locally

361: homogeneous, then in general it belongs to $K$. That is, it is plausible

362: that in 4D, $CSI \equiv H \cup \KC$.

363: 

364: In higher dimensions it was also shown that $\ell$ is geodesic,

365: non-expanding, shear-free and non-twisting (i.e., $L_{ij}=0$) in

366: $VSI$ spacetimes \cite{Higher}. In locally homogeneous spacetimes, in general 

367: there exists a null frame in which the

368: $L_{ij}$ are constants. We therefore anticipate that in $CSI$

369: spacetimes that are not locally homogeneous, $L_{ij}=0$. For

370: higher dimensional $CSI$ spacetimes with $L_{ij}=0$, the Ricci and

371: Bianchi identities appear to be identically satisfied

372: \cite{bianchi}. A higher dimensional spacetime which admits a null

373: vector $\ell$ which is geodesic, non-expanding, shear-free and

374: non-twisting, will be denoted as a higher-dimensional Kundt

375: spacetime.

376: 

377: Let us assume that the spacetime admits a geodesic, non-twisting, non-expanding,

378: shear-free null vector ${\ell}$. It can be then shown that the existence of a twist-free and geodesic null vector implies that there exists a local coordinate system $(u,v,x^i)$ such that

379: \beq

380: \d s^2=2\d u\left(\d

381: v+H\d u+W_{ i}\d x^i\right)+\tilde{g}_{ij}(v,u,x^k)\d x^i\d x^j.

382: \eeq

383: In essence, the twist-free condition implies that the vector field is 'surface forming'

384: so that there exists locally an exact null one-form $\d u$.

385: Using a coordinate transformation and a null-rotation we can now bring the metric into

386: the above form. In this coordinate system, $\ell=\frac{\partial}{\partial v}$.

387: In addition, requiring that this vector field is non-expanding and shear-free implies

388: that $\tilde{g}_{ij,v}=0$.

389: 

390: We will therefore consider metrics of the from

391: \beq \d s^2=2\d u\left(\d

392: v+H\d u+W_{ i}\d x^i\right)+\tilde{g}_{ij}(u,x^k)\d x^i\d x^j,

393: \label{Kundt}

394: \eeq

395: where  $H=H(v,u,x^k)$ and

396: $W_{i}=W_{i}(v,u,x^k)$. The metric (\ref{Kundt})

397: possesses a null vector field $\ell$ obeying \beq

398: \ell^{A}\ell_{B;A}=\ell^{A}_{~;A}=\ell^{A;B}\ell_{(A;B)}=\ell^{A;B}\ell_{[A;B]}=0;

399: \eeq i.e., it is geodesic, non-expanding, shear-free and

400: non-twisting. We will refer to the metrics (\ref{Kundt}) as

401: higher-dimensional \emph{Kundt metrics} (or simply Kundt metrics),

402: since they generalise the 4-dimensional Kundt metrics.

403: 

404: 

405: \subsection{Lorentzian {$CSI$}} All examples to the authors knowledge

406: of Lorentzian {$CSI$} spacetimes are of the following two forms:

407: \begin{enumerate}

408:  \item{}Homogeneous spaces

409:  \item{}A subclass of the Kundt spacetimes

410: \end{enumerate}

411: For the homogeneous spaces, their decomposition can be of the most general type,

412: \[ \nabla^{(k)}R= \sum_{b=-(2+k)}^{2+k}\left(\nabla^{(k)}R\right)_b, \]

413: since a frame can be chosen such that all components are constants. However, for all known examples of non-homogeneous {$CSI$} metrics, a frame can be found such that we have the decomposition

414: \[  \nabla^{(k)}R= \sum_{b=-(2+k)}^{0}\left(\nabla^{(k)}R\right)_b, \]

415: with constant boost-weight zero components. 

416: 

417: Since a homogeneous space is automatically a {$CSI$} space, we

418: will in this paper consider metrics which are not necessarily

419: homogeneous. Since all known examples of non-homogeneous metrics

420: are of Kundt form, we will henceforth only consider Kundt metrics.

421: 

422: \section{Warped product {$CSI$}}

423: 

424: It is known that  $\RC$ spacetimes can be

425: constructed in all dimensions via warped products. Let us first

426: determine necessary and sufficient conditions so that the warped

427: product of a homogeneous space and a $VSI$ is $CSI$.

428: 

429: We consider the warped product metric $g_{AB}=\stackrel{1}{g}_{ab}

430: \oplus\; e^{2\tau}\hspace{-0.1cm}\stackrel{2}{g}_{\alpha\beta}$

431: where $\stackrel{1}{g}$ is a Riemannian homogenous space and

432: $\stackrel{2}{g}$ is a $k$-dimensional {$VSI$} Lorentzian

433: manifold, the warping function $\tau$ only depends on points of

434: the homogenous space.  Using the left-invariant

435: frame of the homogeneous space\footnote{Hatted indices are the

436: preferred null-frame indices.},

437: 

438: \begin{equation}

439: {\sf e}_{\hat a}=\left\{{\bf m}_{\hat k},\ldots,{\bf m}_{\widehat{N-1}}\right\},

440: \end{equation}

441: we complete it to a basis of the warped product by appending the null frame,

442: 

443: \begin{equation}

444: {\sf e}_{\hat\alpha}=\left\{ {\bf \ell}=e^{-\tau}{\bf \ell'},{\bf n}=e^{-\tau}{\bf n'},{\bf m}_{\hat 2}=e^{-\tau}{\bf

445: m'}_{\hat 2},\ldots,{\bf m}_{\widehat{k-1}}=e^{-\tau}{\bf m'}_{\widehat{k-1}} \right\},

446: \end{equation}

447: where the primed vectors represent the canonical {$VSI$} frame in which ${\bf \ell'}$ (and hence $\ell$) is the aligned

448: geodesic null congruence that is expansion, shear, and twist free. The non-vanishing inner products are such that

449: 

450: \begin{equation}

451: g_{AB}=2\ell_{(A}n_{B)}+\delta_{\hat\mu\hat\nu}m^{\hat\mu}_{~A}m^{\hat\nu}_{~B}+\delta_{\hat m\hat n}m^{\hat m}_{~A}m^{\hat n}_{~B}.

452: \end{equation}

453: 

454: Defining $T_{\hat m\hat p}=\tau_{,\hat m ; \hat p}+\tau_{,\hat m}\tau_{,\hat p}$, we shall show that if $\tau_{,m}\tau^{,m}$ and $T_{\hat m

455: \hat p}$ are constant then this implies {$CSI$}.  These two conditions imply that $\tau^{,\hat m}$ is an eigenvector of $T_{\hat m

456: \hat p}$ with constant eigenvalue $\tau_{,m}\tau^{,m}$, and thus $\tau_{,\hat b}$ must be constant.

457: 

458: Following \cite{Higher}, we perform a decomposition of the Riemann

459: tensor for the warped product to obtain

460: 

461: \begin{eqnarray}

462: R_{ABCD} = 4R_{\hat 0\hat 1\hat 0\hat 1}n_{\{A}\ell_{B}n_{C}\ell_{D\}}+ 8R_{\hat 0\hat \mu\hat 1\hat\lambda}n_{\{A}m^{\hat \mu}_{~

463: B}\ell_{C}m^{\hat \lambda}_{~ D\}}+

464: 8R_{\hat 0\hat a\hat 1\hat b}n_{\{A}m^{\hat a}_{~B}\ell_{C}m^{\hat b}_{~D\}} \nonumber \\

465: +R_{\hat \mu\hat \nu\hat \lambda\hat \sigma}m^{\hat \mu}_{~\{A}m^{\hat\nu}_{~B}m^{\hat\lambda}_{~ C}m^{\hat\sigma}_{~D\}}

466: +4R_{\hat a\hat\mu\hat b\hat\lambda}m^{\hat a}_{~\{A}m^{\hat\mu}_{~ B}m^{\hat b}_{~ C}m^{\hat\lambda}_{~ D\}} \label{riem}\\

467: +R_{\hat a\hat b\hat c\hat d}m^{\hat a}_{~ \{A}m^{\hat b}_{~ B}m^{\hat c}_{~ C}m^{\hat d}_{ D\}}+\text{negative boost weight terms}\cdots  \nonumber

468: \end{eqnarray}

469: Evidently the Riemann tensor is of boost order zero with boost weight zero components:

470: 

471: \begin{eqnarray}

472: R_{\hat 0\hat 1\hat 0\hat 1} &=& \tau_{,m}\tau^{,m}       \label{rbw0.1} \\

473: R_{\hat 0\hat \mu\hat 1\hat \lambda} &=& -\tau_{,m}\tau^{,m}\stackrel{2}{\delta}_{\hat \mu\hat\lambda}  \label{rbw0.2}  \\

474: R_{\hat 0\hat a\hat 1\hat b} &=&  -T_{\hat a\hat b}                       \label{rbw0.3} \\

475: R_{\hat\mu\hat\nu\hat\lambda\hat\sigma} &=& \tau_{,m}\tau^{,m}

476: \left[\stackrel{2}{\delta}_{\hat\nu\hat\lambda}\stackrel{2}{\delta}_{\hat\mu\hat\sigma} -

477: \stackrel{2}{\delta}_{\hat\mu\hat\lambda}\stackrel{2}{\delta}_{\hat\nu\hat\sigma} \right]  \label{rbw0.4} \\

478: R_{\hat a\hat\mu\hat b\hat\lambda} &=& - T_{\hat a\hat b}\stackrel{2}{\delta}_{\hat\mu\hat\lambda} \label{rbw0.5} \\

479: R_{\hat a\hat b\hat c\hat d} &=& \stackrel{1}{R}_{\hat a\hat b\hat c\hat d}.  \label{rbw0.6}

480: \end{eqnarray}

481: The negative boost weight components, arising from the {$VSI$} spacetime, are given by

482: \begin{eqnarray}

483: R_{\hat 1\hat 0\hat 1\hat \mu} &=& e^{-2\tau}\stackrel{2}{R}_{\hat 1\hat 0\hat 1\hat \mu}  \label{rbw-1.4}  \\

484: R_{\hat 1\hat\mu\hat\nu\hat\lambda} &=& e^{-2\tau}\stackrel{2}{R}_{\hat 1\hat\mu\hat\nu\hat\lambda}  \label{rbw-1.5}  \\

485: R_{\hat 1\hat\mu\hat 1\hat\nu} &=& e^{-2\tau}\stackrel{2}{R}_{\hat 1\hat\mu\hat 1\hat\nu}.  \label{rbw-2.6}

486: \end{eqnarray}

487: The condition that $\tau_{,\hat b}$ is constant (and hence $\tau_{,m}\tau^{,m}$ and 

488: $T_{\hat m\hat p}$ are constant) implies that

489: the boost weight zero components of the Riemann tensor are constant; the absence of positive boost weight components

490: then gives {$CSI$}$_{0}$.

491: 

492: The covariant derivative of the null frame has the form,

493: \begin{eqnarray}

494: \ell_{K;E} &=& -\tau_{,\hat b}m^{\hat b}_{~ K}\ell_{E}+\gamma_{\hat\lambda\hat 0\hat 1}m^{\hat\lambda}_{~ K}\ell_{E}

495: +\gamma_{\hat 1\hat 0\hat \lambda}\ell_{K}m^{\hat\lambda}_{~E} + \cdots \label{dl} \\ %\gamma_{(1)(0)(1)}\ell_{K}\ell_{E} \\

496: n_{K;E} &=& -\tau_{,\hat b}m^{\hat b}_{~ K}n_{E}-\gamma_{\hat 1\hat 0\hat \lambda}n_{K}m^{\hat \lambda}_{~ E} + \cdots \label{dn}\\

497: %\gamma_{(\mu)(1)(\lambda)}m^{(\mu)}_{\quad K}m^{(\lambda)}_{\quad E}-\gamma_{(1)(0)(1)}n_{K}\ell_{E}

498: %+\gamma_{(\lambda)(1)(1)}m^{(\lambda)}_{\quad K}\ell_{E} \\

499: m_{\hat \lambda K;E} &=& -\tau_{,\hat b}m^{\hat b}_{~K}m_{\hat \lambda E}+\gamma_{\hat \mu\hat \lambda\hat \nu}m^{\hat\mu}_{~

500: K}m^{\hat \nu}_{~E} - \gamma_{\hat \lambda\hat 0\hat 1}n_{K}\ell_{E}+ \cdots \label{dmv}\\

501: %\gamma_{(\mu)(\lambda)(1)}m^{(\mu)}_{\quad K}\ell_{E}

502: %-\gamma_{(\lambda)(1)(\mu)}\ell_{K}m^{(\mu)}_{\quad E}-\gamma_{(\lambda)(1)(1)}\ell_{K}\ell_{E} \\

503: m_{\hat a K;E} &=& \tau_{,\hat a}\ell_{K}n_{E}+\tau_{,\hat a}n_{K}\ell_{E} +\tau_{,\hat a}\delta_{\hat \mu\hat\lambda}m^{\hat\mu}_{~

504: K}m^{\hat\lambda}_{~ E}+\stackrel{1}{\gamma}_{\hat b\hat a\hat c}m^{\hat b}_{~K}m^{\hat c}_{~ E} \label{dma}

505: \end{eqnarray}

506: where $\cdots$ denotes terms of lower order boost weight with

507: respect to previous terms in an expression, and the unspecified

508: $\gamma$'s correspond to {$VSI$} Ricci-rotation coefficients

509: \cite{Higher} [for example, $\gamma_{\hat 1\hat 0\hat

510: \lambda}=e^{-\tau}L_{\hat 1\hat\lambda}$], except for the

511: $\stackrel{1}{\gamma}$ which are the rotation coefficients

512: associated with the homogeneous space. Equations

513: (\ref{dl})-(\ref{dma}) have leading order boost weight terms of

514: $+1,~-1,~0$ and $0$, respectively, which is the same boost weight

515: as the corresponding null frame vector before taking a covariant

516: derivative.  It follows that the covariant derivative of the

517: $\{\cdot\}$ quantities appearing in (\ref{riem}) will produce

518: terms of equal or lesser boost weight.  Furthermore, in

519: \cite{Higher} it was proven that the negative boost weight

520: curvature components appearing in (\ref{riem}) will remain

521: negative upon covariant differentiation.  Therefore, if the boost

522: weight zero components of the Riemann tensor and its derivatives

523: are constant then any number of covariant derivatives of the

524: Riemann tensor will be of boost order zero, which implies that the

525: warped product will be {$CSI$}.

526: 

527: Supposing that the boost weight zero components of $\nabla^{(n)}R$ are constant, then for $\nabla^{(n+1)}R$ the boost

528: weight zero components will have contributions from $\tau_{,\hat a}$, which is constant, and the non-constant $\gamma$'s

529: arising from the {$VSI$} spacetime, however these boost weight zero {$VSI$} 

530: $\gamma$'s do not occur as a result of

531: Theorem 5.1.  Therefore we have that if $\tau_{,\hat b}$ is constant, then the warped product is {$CSI$}.

532: 

533: For reference, we  include the covariant derivative of the Riemann tensor.  By using the Ricci identity we have,

534: \begin{equation}

535: \tau^{,\hat a}\stackrel{1}{R}_{\hat a\hat b\hat c\hat d}+2T_{\hat b[\hat d}\tau_{,\hat c]}=T_{\hat b\hat c;\hat d}-T_{\hat b\hat d;\hat c},   \label{riciden}

536: \end{equation}

537: which can be used to express the non-vanishing boost weight zero components of the covariant derivative of the Riemann

538: tensor as

539: \begin{eqnarray}

540: R_{\hat 1\hat b\hat c\hat d;\hat 0} &=& T_{\hat b\hat c;\hat d}-T_{\hat b\hat d;\hat c}=R_{\hat 0\hat b\hat c\hat d;\hat 1} \label{drbw0.1} \\

541: R_{\hat \mu\hat b\hat c\hat d;\hat \lambda} &=& \left[T_{\hat b\hat c;\hat d}-T_{\hat b\hat d;\hat c}\right]\delta_{\hat\mu\hat\lambda} \label{drbw0.2} \\

542: R_{\hat 0\hat b\hat 1\hat c;\hat e} &=& -T_{\hat b\hat c;\hat e}   \label{drbw0.3} \\

543: R_{\hat b\hat \mu\hat c\hat \lambda;\hat e} &=& -T_{\hat b\hat c;\hat e}\delta_{\hat \mu\hat \lambda}.   \label{drbw0.4}

544: \end{eqnarray}

545: As expected all components in (\ref{drbw0.1})-(\ref{drbw0.4}) are constant (as a result of requiring $\tau_{,\hat b}$ to

546: be constant).

547: 

548: We shall now show the converse; that is, if the Ricci invariants are constant then $\tau_{,\hat b}$ is constant.

549: In local coordinates, the Ricci tensor and Ricci scalar are

550: \begin{eqnarray}

551: R_{\mu}^{~\rho} &=& e^{-2\tau}\stackrel{2}{R}_{\mu}^{~\rho}

552: -[T+(k-1)\tau_{,m}\tau^{,m}]\delta_{\mu}^{\rho} \label{ric2} \\

553: R_{m}^{~ p} &=& \stackrel{1}{R}^{~ p}_{m}-kT_{m}^{~ p}  \label{ric1} \\

554: R &=&\stackrel{1}{R}-k[2T+(k-1)\tau_{,m}\tau^{,m}],\label{ricsc}

555: \end{eqnarray}

556: where in (\ref{ricsc}) we have used that $\stackrel{2}{R}=0$ and set $T=\Tr(T_{m}^{~p})$. Choosing new coordinates

557: $\tilde{x}^{1}=\tau(x^{m}), \tilde{x}^{2}=x^{2},\ldots$ so 

558: that $\tau_{,m}\tau^{,m}=\tilde{g}^{11}$, then

559: 

560: \begin{equation}

561: \tilde{T}_{mn}= -\tilde{\Gamma}^{1}_{mn}+\delta^{1}_{m}\delta^{1}_{n}.

562: \end{equation}

563: 

564: Now, with normal coordinates along $\tilde{x}^{1}$ we have $\tilde{g}^{11}=1$.  Therefore from (\ref{ricsc}), since

565: $\stackrel{1}{R}$ is constant we have that $\tilde{T}$, and hence $T$, is constant.  In this coordinate system the

566: constraint $\tilde{T}$ constant implies that $\tilde{g}^{mn}\tilde{g}_{mn,1}$ is constant.  Constant Ricci scalar

567: implies that $T$ and $\tau_{,m}\tau^{,m}$ are constant. Further conditions are then provided by the constancy of higher

568: degree Ricci invariants.

569: 

570: From (\ref{ric2}) and (\ref{ric1}),  the Ricci tensor of a warped product is $R_{M}^{\ \ N}=R_{m}^{\ \ n} \oplus

571: R_{\mu}^{\ \ \nu}$, thus a degree $d$ Ricci invariant always has the form

572: 

573: \begin{equation}

574: R_{N_{1}}^{\ \ N_{2}}R_{N_{2}}^{\ \ N_{3}}\cdots R_{N_{d}}^{\ \ N_{1}}=R_{n_{1}}^{\ \ n_{2}}R_{n_{2}}^{\ \ n_{3}}\cdots

575: R_{n_{d}}^{\ \ n_{1}} + R_{\nu_{1}}^{\ \ \nu_{2}}R_{\nu_{2}}^{\ \ \nu_{3}}\cdots R_{\nu_{d}}^{\ \ \nu_{1}}.

576: \label{ricinv}

577: \end{equation}

578: The second term of (\ref{ricinv}) is constant since $T$ and $\tau_{,m}\tau^{,m}$ are constant; hence invariants

579: constructed from (\ref{ric1}) must be constant. Assuming the determinant of $R_{m}^{ ~p}$ or $\stackrel{1}R^{~

580: p}_{m}$ is nonzero, then there exists a frame in which $R_{m}^{~ p}$ and $\stackrel{1}R^{~ p}_{m}$ can

581: simultaneously be expressed in block diagonal form with constant eigenvalues \cite{Petrov}. Therefore, (\ref{ric1}) implies that in

582: this frame $T_{m}^{~ p}$ must also have constant components and thus constant invariants. 

583: In this case, the constancy of the Ricci

584: invariants completely determines the conditions on $T_{m}^{~ p}$ and specifies a frame in which the

585: components of each term in (\ref{ric1}) are constant. The frame in which $\stackrel{1}R^{~ p}_{m}$ and $T_{m}^{~ p}$ are

586: constant is determined by the left-invariant 1-forms.  As before, the conditions that $T_{\hat a\hat b}$ and

587: $\tau_{,m}\tau^{,m}$ are constant imply $\tau_{,\hat a}$ is constant, and consequently 

588: constant Ricci invariants implies {$CSI$}.

589: 

590: 

591: Many examples of {$CSI$} spacetimes can be found among the cases where the homogeneous space is a solvable Lie group\footnote{Recall that a Lie group can be equipped with a left-invariant metric (i.e., invariant under the left action) in the standard way, see e.g. \cite{Milnor}.}:

592: \begin{thm}

593: Assume that $\mathcal{M}_H$ is a solvable Lie group equipped with a left-invariant metric. Assume also that $\mathcal{M}_H$ is connected and simply connected (i.e., $H^1(\mathcal{M}_H)=0$). Then there exists a non-constant $\tau$ such that the warped product of $\mathcal{M}_H$ and a $VSI$, as constructed above, is a {$CSI$}.

594: \end{thm}

595: \begin{proof}

596: We identify $\mathcal{M}_H=G$, where $G$ is a solvable Lie group, with the Lie algebra $\mf{g}=T_{e}G$. Since $\mf{g}$ is solvable the set $[\mf{g},\mf{g}]$ will be a proper vector subspace of $\mf{g}$. Thus for a solvable Lie algebra, there exists a non-zero ${\bf X}\in\mf{g}\cap[\mf{g},\mf{g}]^{\perp}$ where $[\mf{g},\mf{g}]^{\perp}$ is the complement of the derived algebra $[\mf{g},\mf{g}]$. This further implies that there is a left-invariant one-form ${\mbold\omega}$ such that $\d {\mbold\omega}=0$. Since $H^1(\mathcal{M}_H)=0$, there exists a non-constant function $\phi$ such that ${\mbold\omega}=\d \phi$. We can now choose $\tau=p\phi$ for a constant $p$.

597: \end{proof}

598: 

599: In 2-dimensions flat space is the only {$VSI$} metric, and in

600: the appendix we list all of the 3-dimensional {$VSI$} metrics.

601: Examples of {$CSI$} warped products where the homogeneous space

602: is either $\mathbb{E}^n$ or $\mathbb{H}^n$ yield a linear function

603: for $\tau$ (since both of these can be considered as

604: solvmanifolds); however, for $\mathbb{S}^n$, since the coordinates

605: are cyclic, we find that $\tau$ is a constant. Another example, in

606: 5-dimensions, is where the homogeneous space is a Bianchi type VIII Lie group ($\cong SL(2,\mathbb{R})$)

607: warped with a 2-dimensional {$VSI$}, with coordinates $u,v,x,y,z$

608: the metric is

609: \beq

610: \d s^2 &=& e^{2\tau}\left[2\d u(\d v+H\d u)\right]\nonumber \\

611: && +\frac{a^2}{y^2}(\cos z\d x+\sin z\d y)^2+\frac{b^2}{y^2}(\cos z\d y-\sin z\d x)^2+c^2\left(\d z+\frac{\d x}{y}\right)^2.

612: \eeq

613: In this case $H(u,v)$ must be linear in $v$ to give {$CSI$}, and it turns out that for the warped product to be {$CSI$} we have to consider two separate cases:

614: \begin{enumerate}

615: \item{} The maximal isometry group of the homogeneous space is 3-dimensional ($a\neq b$): $\tau$ is a constant.

616: \item{} The maximal isometry group of the homogeneous space is 4-dimensional ($a=b$): $\tau$ can be non-constant, $\tau=\alpha \ln y+\beta z$.

617: \end{enumerate}

618: The reason why we have to distinguish these to cases is that in the first case, the isometry group is the semi-simple group, $SL(2,\mathbb R)$. However, in the

619: second case, the isometry group allows for a transitive group

620: which is solvable (Bianchi type III).

621: 

622: \section{Kundt {$CSI$} metrics}

623: 

624: 

625: Let us next consider the higher-dimensional Kundt metrics (\ref{Kundt})

626: in the form

627: \beq \d

628: s^2=2\d u\left(\d v+H\d u+W_{\hat i}{\bf m}^{\hat

629: i}\right)+\delta_{\hat i\hat j}{\bf m}^{\hat i}{\bf m}^{\hat j}.

630: \label{Kundtmetric}\eeq 

631: It is convenient to introduce the null frame

632: \beq

633: \ell&=& \d u, \\

634: {\bf n}&=& \d v+H\d u+W_{\hat i}{{\bf m}}^{\hat i}, \\

635: {\bf m}^{\hat i}&=& {\sf e}^{\hat i}_{~j}(u; x^{k})\d x^j,\\

636: (\eta_{AB}) &=&\begin{bmatrix}

637: 0 & 1 & 0 \\

638: 1 & 0 & 0 \\

639: 0 & 0 & \delta_{\hat i\hat j}

640: \end{bmatrix}

641: \eeq

642: such that

643: \beq

644: \d S^2_H\equiv\delta_{\hat i\hat j}{\bf m}^{\hat i}{\bf m}^{\hat j}=\tilde{g}_{ij}\d x^i\d x^j,\quad \tilde{g}_{ij,v}=0.

645: \eeq

646: 

647: There is a class of coordinate transformations that preserve the form of the metric (\ref{Kundtmetric}). In particular, we can define new variables,

648: \beq

649: (v',u',x'^i)=(v,u,f^i(u;x^k)).

650: \eeq

651: For our purposes we can always use this coordinate transformation to simplify the spatial metric $\tilde{g}_{ij}$:

652: \begin{thm}

653: Consider the metric (\ref{Kundt}) and assume that there exists a null frame $\left\{\ell_A,n_A,m^{\hat{i}}_{~A}\right\}$ such that all scalars $R_{\hat{i}\hat{j}\hat{k}\hat{l}}\equiv R_{ABDC}m^{~A}_{\hat{i}}m^{~B}_{\hat{j}}m^{~C}_{\hat{k}}m^{~D}_{\hat{l}}$ and $R_{\hat{i}\cdots;\hat{j}_1\cdots\hat{j}_n}\equiv R_{A\cdots ;C_1\cdots C_n }m^{~A}_{\hat{i}}\cdots m^{~C_1}_{\hat{j}_1}\cdots m^{~C_n}_{\hat{j}_n}$ are constants. Then there exists (locally) a coordinate transformation $(v',u',x'^i)=(v,u,f^i(u;x^k))$ such that

654: \beq

655: \tilde{g}_{ij}\equiv \tilde{g}'_{kl}\frac{\partial f^k}{\partial x^i}\frac{\partial f^l}{\partial x^j}, \quad \tilde{g}'_{ij,u'}=0.

656: \eeq

657: Moreover, $\d S_H^2=\tilde{g}'_{ij}\d x'^i\d x'^j$ is a locally homogeneous space.

658: \label{thm:Kundt}

659: \end{thm}

660: \begin{proof}

661: By calculating the curvature tensor of the metric (\ref{Kundt}), and its covariant derivatives, we note that

662: \beq

663: R_{\hat{i}\hat{j}\hat{k}\hat{l}}=\widetilde{R}_{\hat{i}\hat{j}\hat{k}\hat{l}}, \quad R_{\hat{i}\cdots;\hat{j}_1\cdots\hat{j}_n}=\widetilde{R}_{\hat{i}\cdots;\hat{j_1}\cdots\hat{j}_n},

664: \eeq

665: where the tilde refers to curvature tensors with respect to the spatial metric $\tilde{g}_{ij}(u,x^k)$. Since this metric is Riemannian with all its components constant we can use the results of \cite{PTV} which state that the metrics $\tilde{g}_{ij}(u,x^k)$, for different $u$, are equivalent up to isometries. Thus, given $u_0$, there exists for every $u$ sufficiently close to $u_0$ an isometry $\psi_u(x^k)$ such that $(\psi_u^*\tilde{g})(u)=\tilde{g}(u_0)$. Since, we assume that $\tilde{g}$ is sufficiently smooth in a neighbourhood of $u_0$, we can find a sufficently smooth $\psi_u$ in $u$. The map $\psi_u(x^k)$ provides us with the functions $f^i$ in the theorem by composition  with the coordinate charts. Hence, we have $\tilde{g}'=\tilde{g}\big|_{u=u_0}$. Moreover, $\tilde{g}'$ is a locally homogeneous space \cite{PTV}.

666: \end{proof}

667: 

668: 

669: Henceforth, we shall assume that we consider solutions in the set $\FC  \bigcap \KC$.

670: Consequently, from Theorem 4.1, there is no loss of generality in assuming that the metric 

671: (\ref{Kundt}) has $\tilde{g}_{ij,u}=0$. 

672: The remaining coordinate freedom preserving the Kundt form is then:

673: \begin{enumerate}

674: \item{} $(v',u',x'^i)=(v,u,f^i({x}^k))$ and $J^i_{~j}\equiv \frac{\partial f^i}{\partial x^j}$.

675: \beq

676: H'= H, \quad W'_i=W_j\left(J^{-1}\right)^j_{~i}, \quad \tilde{g}'_{ij}=\tilde{g}_{kl}\left(J^{-1}\right)^k_{~i}\left(J^{-1}\right)^l_{~j}.\nonumber

677: \eeq

678: \item{} $(v',u',x'^i)=(v+h(u,x^k),u,x^i)$

679: \beq

680: H'=H-h_{,u}, \quad W'_i=W_i-h_{,i}, \quad \tilde{g}_{ij}'=\tilde{g}_{ij}.\nonumber

681: \eeq

682: \item{} $(v,u,x^i)=(v/g_{,u}(u),g(u),x^i)$

683: \beq

684: H'=\frac 1{g_{,u}^2}\left(H+v\frac{g_{,uu}}{g_u}\right), \quad W'_i=\frac 1{g_{,u}}W_i, \quad \tilde{g}'_{ij}=\tilde{g}_{ij}.\nonumber

685: \eeq

686: \end{enumerate}

687: 

688: 

689: 

690: \subsection{ {$CSI$}$_0$ spacetimes}

691: The linearly independent components of the Riemann tensor with boost weight 1 and 0 are:

692: \beq

693: R_{\hat 0\hat 1\hat 0\hat i}&=&-\frac 12 W_{\hat i,vv}, \\

694: R_{\hat 0\hat 1\hat 0\hat 1}&=& -H_{,vv}+\frac 14\left(W_{\hat{i},v}\right)\left(W^{\hat i,v}\right), \\

695: R_{\hat 0\hat 1\hat i\hat j}&=& W_{[\hat i}W_{\hat j],vv}+W_{[\hat i;\hat j],v}, \\

696: R_{\hat 0\hat i\hat 1\hat j}&=& \frac 12\left[-W_{\hat j}W_{\hat i,vv}+W_{\hat i;\hat j,v}-\frac 12 \left(W_{\hat i,v}\right)\left(W_{\hat j,v}\right)\right], \\

697: R_{\hat i\hat j\hat n\hat m}&=&\tilde{R}_{\hat i\hat j\hat n\hat m}.

698: \eeq

699: Hence, the spacetime is a {$CSI$}$_0$ spacetime if there exists a frame $\left\{ \ell, {\bf n}, {\bf m}^{\hat i}\right\}$, a constant $\sigma$, anti-symmetric matrix ${\sf a}_{\hat i\hat j}$, and symmetric matrix ${\sf s}_{\hat i\hat j}$ such that

700: \beq

701: W_{\hat i,vv} &=& 0, \label{Wcsi1}\\

702: H_{,vv}-\frac 14\left(W_{\hat i,v}\right)\left(W^{\hat i,v}\right) &=& \sigma, \\

703: W_{[\hat i;\hat j],v} &=& {\sf a}_{\hat i\hat j}, \\

704: W_{(\hat i;\hat j),v}-\frac 12 \left(W_{\hat i,v}\right)\left(W_{\hat j,v}\right) &=& {\sf s}_{\hat i\hat j},

705: \label{Wcsi4}\eeq

706: and the components $\tilde{R}_{\hat i\hat j\hat m\hat n}$ are all constants (i.e., $\d S^2_H$ is curvature homogeneous).

707: 

708: We note that the first equation implies that

709: \beq

710: W_{\hat i}(v,u,x^k)=v{W}_{\hat i}^{(1)}(u,x^k)+{W}_{\hat i}^{(0)}(u,x^k),

711: \eeq

712: while the second implies

713: \beq

714: H(v,u,x^k)=\frac{v^2}{8}\left[4\sigma+({W}_i^{(1)})({W}^{(1)i})\right]+v{H}^{(1)}(u,x^k)+{H}^{(0)}(u,x^k).

715: \eeq

716: If $\d S^2_H=\delta_{\hat i\hat j}{\bf m}^{\hat i}{\bf m}^{\hat j}$ is a (locally) homogeneous

717: Riemannian space, then there exists a frame where  $\tilde{R}_{\hat i\hat j\hat m\hat n}$ are all

718: constants. In general, curvature homogeneous does not imply homogeneous; however, 

719: since we are mostly interested in the {$CSI$} case, in light of Theorem \ref{thm:Kundt},  

720: we will henceforth take $\d S^2_H$ as a (locally) homogeneous metric.

721: 

722: \subsection{ {$CSI$}$_1$ spacetimes}

723: For a {$CSI$}$_0$ spacetime, we have

724: \beq

725: R_{\hat 0\hat 1\hat 0\hat i;\hat 1}&=&-\frac 12\left[\sigma W_{\hat i,v}-\frac 12({\sf s}_{\hat j\hat i}+{\sf a}_{\hat j\hat i})W^{\hat j,v}\right], \\

726: R_{\hat 0\hat i\hat j\hat k;\hat 1}&=&-\frac 12\left[W^{\hat n,v}\tilde{R}_{\hat n\hat i\hat j\hat k}-W_{\hat i,v}{\sf a}_{\hat j\hat k}+({\sf s}_{\hat i[\hat j}+{\sf a}_{\hat i[\hat j})W_{\hat k],v}\right].

727: \eeq

728: For the spacetime to be {$CSI$}$_1$, it is sufficient to require that the above components are constants; i.e.

729: \beq

730: \sigma W_{\hat i,v}-\frac 12({\sf s}_{\hat j\hat i}+{\sf a}_{\hat j\hat i})W^{\hat j,v} &=& {\mbold\alpha}_{\hat i}, \\

731: \left({\sf s}_{\hat i\hat j}+{\sf a}_{\hat i\hat j}\right)_{;\hat k}-\left({\sf s}_{\hat i\hat k}+{\sf a}_{\hat i\hat k}\right)_{;\hat j}&=& {\mbold\beta}_{\hat i\hat j\hat k},

732: \eeq

733: where the Ricci identity has been used to rewrite the latter condition.

734: 

735: 

736: \section{Examples of {$CSI$} spacetimes of Kundt form}

737: Various classes of {$CSI$} spacetimes 

738: that arise as members of $\FC \bigcap \KC$

739: can now be found. The solutions come in classes according to the properties of $W^{(1)}_{\hat{i}}(u,x^k)$.  In the following, $\mathcal{M}_H$ is a locally homogeneous Riemannian space.

740: \subsection{$W^{(1)}_{\hat{i}}(u,x^k)=0$}

741: Assuming $W^{(1)}_{\hat{i}}(u,x^k)=0$ immediately makes the metric (\ref{Kundtmetric}) a {$CSI$} space.

742: 

743: A special subcase of this class which is worth mentioning is the Brinkmann metrics for which $H$ and $W_i$ are all independent of $v$; thus $\sigma={\sf s}_{ij}={\sf a}_{ij}=0$. In this case, the expressions for the curvature tensors simplify drastically. In particular, if $F_{\hat{j}\hat{i}}\equiv 2W_{[\hat{i};\hat{j}]}$, the Ricci tensor is

744: \beq

745: R_{\hat{1}\hat{1}}=\tilde{\Box}H-\frac 14\tilde{F}^2, \quad R_{\hat{1}\hat{i}}=\tilde{\nabla}^{\hat{j}}F_{\hat{j}\hat{i}}, \quad R_{\hat{i}\hat{j}}=\tilde{R}_{\hat{i}\hat{j}}.

746: \eeq

747: 

748: \subsection{$W^{(1)}_{\hat{i}}(u,x^k)=\mathrm{constant}$} This case is {$CSI$} if and only if the simpler metric, $\widetilde{\mathcal{M}}$:

749: \beq

750: \widetilde{\d s^2}=2\d u\left(\d v+\frac{v^2}{2}\tilde{\sigma}\d u+vW_{\hat i}^{(1)}{\bf m}^{\hat i}\right)+\delta_{\hat i\hat j}{\bf m}^{\hat i}{\bf m}^{\hat j},

751: \eeq

752: is a homogeneous space. The curvature invariants of the Kundt metrics in this class will have the same invariants as $\widetilde{\mathcal{M}}$; i.e. $\mathcal{I}=\mathcal{I}(\widetilde{\mathcal{M}})$.

753: 

754: The homogeneous space $\mathcal{M}_H$ can be considered as a quotient space  $\mathcal{M}_H=G/H$, 

755: where $G$ is the isometry group and $H$ is the isotropy group. The corresponding 

756: Lie algebra decomposition is $\mf{g}=\mf{h}+\mf{m}$, and we let this correspond to the algebra of left-invariant vectors. We note that if there exists a subalgebra $\mf{k}\subset\mf{m}$ such that $[\mf{k}^{\perp},\mf{k}]=0$, where $\mf{k}^{\perp}$ is the complement of $\mf{k}$ in $\mf{g}$, then we can choose an $\mathrm{Ad}(H)$-invariant ${\bf W}$:

757: \[ {\bf W}\in \mf{k}.\]

758: By letting  $W^{(1)}_{\hat{i}}{\bf m}^{\hat{i}}$ be its corresponding left-invariant one-form, the associated Kundt spacetime will be a {$CSI$}.

759: 

760: Note that some homogeneous spaces may be represented by inequivalent Lie algebra decompositions. 

761: For example, hyperbolic space, $\mb{H}^n$, can both be considered as the quotient $SO(1,n)/SO(n)$ 

762: and as a solvable Lie group $G$ \cite{Hervik}. In the latter case the decomposition gives $\mf{g}=\mf{m}$ and $\mf{h}=0$; hence, $[\mf{h},\mf{m}]=0$.

763: 

764: \subsection{$W^{(1)}_{\hat{i}}(u,x^k)=\phi_{,\hat{i}}=\mathrm{non-constant}$}

765: \subsubsection{$\mathcal{I}=\mathcal{I}(\mathrm{dS}\times \mathcal{M}_H)$} \label{subsect:dSM} $\sigma>0$:

766: \beq

767: \d s^2&=&\cos^2(\sqrt{\sigma}x)\left[2\d u\left(\d v+\widetilde{H}\d u+\widetilde{W}_i{\bf m}^i\right)\right]\nonumber \\  && +\d x^2+\frac{1}{\sigma}\sin^2(\sqrt{\sigma}x)\d S^2_{S^n} +\d S^2_{H},

768: \label{eq:csi2}\eeq

769: where

770: \beq

771: && W^{(1)}_{\hat{x}}=2\sqrt{\sigma}\tan(\sqrt\sigma x),\nonumber \\

772: && \widetilde{W}_i = \widetilde{W}_i^{(0)}(u,x^j), \quad

773: \widetilde{H}=\frac{v^2}{2}\sigma+v\widetilde{H}^{(1)}(u,x^j)+\widetilde{H}^{(0)}(u,x^j).

774: \eeq

775: The Riemann tensor is of type:

776: \[ R=(R)_0+(R)_{-1}+(R)_{-2}. \]

777: Special cases:

778: \begin{enumerate}

779: \item{} If $\widetilde{W}^{(0)}_i=\frac{1}{\sigma}\widetilde{H}^{(1)}_{~~,i}$ then $(R)_{-1}=0$.

780: \item{} If $\widetilde{W}^{(0)}_i=\widetilde{H}^{(1)}=\widetilde{H}^{(0)}=0$ 

781: then $(R)_{-1}=(R)_{-2}=0$ and $\mathrm{dS}\times \mathcal{M}_H$.

782: \end{enumerate}

783: 

784: 

785: \subsubsection{$\mathcal{I}=\mathcal{I}(\mathrm{AdS}\times \mathcal{M}_H)$}\label{subsect:AdSM}  $\sigma<0$:

786: \beq

787: \d s^2_I&=&\cosh^2(\sqrt{|\sigma|}x)\left[2\d u\left(\d v+\widetilde{H}\d u+\widetilde{W}_i{\bf m}^i\right)\right]\nonumber \\  && +\d x^2+\frac{1}{|\sigma|}\sinh^2(\sqrt{|\sigma|}x)\d S^2_{S^n} +\d S^2_{H},

788: \label{eq:csi3}\eeq

789: where

790: \beq

791: &&W^{(1)}_{\hat{x}}=-2\sqrt{|\sigma|}\tanh(\sqrt{|\sigma|} x), \nonumber \\

792: &&\widetilde{W}_i = \widetilde{W}_i^{(0)}(u,x^j), \quad

793: \widetilde{H}=-\frac{v^2}{2}|\sigma|+v\widetilde{H}^{(1)}(u,x^j)+\widetilde{H}^{(0)}(u,x^j).

794: \eeq

795: The Riemann tensor is of type:

796: \[ R=(R)_0+(R)_{-1}+(R)_{-2}. \]

797: Special cases:

798: \begin{enumerate}

799: \item{} If $\widetilde{W}^{(0)}_i=-\frac{1}{|\sigma|}\widetilde{H}^{(1)}_{~~,i}$ then $(R)_{-1}=0$.

800: \item{} If $\widetilde{W}^{(0)}_i=\widetilde{H}^{(1)}=\widetilde{H}^{(0)}=0$ 

801: then $(R)_{-1}=(R)_{-2}=0$ and $\mathrm{AdS}\times \mathcal{M}_H$.

802: \end{enumerate}

803: 

804: Let us present another example:

805: \beq

806: \d s^2_{II}&=&\sinh^2(\sqrt{|\sigma|}x)\left[2\d u\left(\d v+\widetilde{H}\d u+\widetilde{W}_i{\bf m}^i\right)\right]\nonumber \\  && +\d x^2+\frac{1}{|\sigma|}\cosh^2(\sqrt{|\sigma|}x)\d S^2_{\mathbb{H}^n} +\d S^2_{H},

807: \label{eq:csi4}\eeq

808: where

809: \beq

810: &&W^{(1)}_{\hat{x}}=2\sqrt{|\sigma|}\coth(\sqrt{|\sigma|} x), \nonumber \\

811: &&\widetilde{W}_i = \widetilde{W}_i^{(0)}(u,x^j), \quad

812: \widetilde{H}=\frac{v^2}{2}|\sigma|+v\widetilde{H}^{(1)}(u,x^j)+\widetilde{H}^{(0)}(u,x^j).

813: \eeq

814: The Riemann tensor is of type:

815: \[ R=(R)_0+(R)_{-1}+(R)_{-2}. \]

816: Special cases:

817: \begin{enumerate}

818: \item{} If $\widetilde{W}^{(0)}_i=\frac{1}{|\sigma|}\widetilde{H}^{(1)}_{~~,i}$ then $(R)_{-1}=0$.

819: \item{} If $\widetilde{W}^{(0)}_i=\widetilde{H}^{(1)}=\widetilde{H}^{(0)}=0$ then $(R)_{-1}=(R)_{-2}=0$.

820: \end{enumerate}

821: 

822: There are more {$CSI$} spacetimes with $\mathcal{I}=\mathcal{I}(\mathrm{(A)dS}\times \mathcal{M}_H)$. They are constructed similarly to the ones above. They are all related to the fact that de Sitter space or Anti-de Sitter space can be written in many ways as fibered spaces over the spheres, $S^n$, or the hyperbolic space, $\mathbb{H}^n$, respectively (see  Appendix \ref{app:4D} for the 4D case).

823: 

824: \subsection{$W^{(1)}_{\hat{a}}(u,x^k)=\mathrm{non-constant}, ~W^{(1)}_{\hat{\alpha}}(u,x^k)=\mathrm{constant}$}

825: This is a generalisation of the Warped {$CSI$} spacetimes considered earlier. The metric can be written

826: \beq

827: \d s^2&=&e^{-2\phi(x^\alpha)}\left[2\d u\left(\d v+\widetilde{H}\d u+\widetilde{W}_a\d x^a+\widetilde{W}_\alpha{\mbold{\omega}}^\alpha\right)+\delta_{ab}\d x^a\d x^b\right] \nonumber \\

828: && +\delta_{\alpha\beta}{\mbold\omega}^\alpha{\mbold\omega}^\beta,

829: \label{eq:csi}\eeq

830: where

831: \beq

832: \widetilde{W}_a &=& v\widetilde{W}_a^{(1)}(u,x^a)+\widetilde{W}_a^{(0)}(u,x^a,x^\alpha), \\

833: \widetilde{W}_\alpha &=& \widetilde{W}_\alpha^{(0)}(u,x^a,x^\alpha), \\

834: \phi &=& \phi(x^\alpha), \quad \phi_{,\alpha}=\mathrm{constant}, \\

835: \widetilde{H}&=&\frac{v^2}{8}(\widetilde{W}^{(1)}_a)(\widetilde{W}^{(1)a})+v\widetilde{H}^{(1)}(u,x^a,x^\alpha)+\widetilde{H}^{(0)}(u,x^a,x^\alpha).

836: \eeq

837: Furthermore, we require that $\delta_{\alpha\beta}{\mbold\omega}^\alpha{\mbold\omega}^\beta=\tilde{g}_{\alpha\beta}(x^\sigma)\d x^\alpha \d x^\beta$ is a 

838: Riemannian homogeneous space, $\mathcal{M}_H$. The allowed forms of the function $\phi$ depend on the  homogeneous space $\mathcal{M}_H$. Topologically, this space is a fibered 

839: manifold with base manifold $\mathcal{M}_H$ and we will require that the fiber is a {$VSI$} space; 

840: in particular, this means that $\widetilde{W}_a$ satisfy the {$VSI$} equations 

841: (\ref{Wcsi1})-(\ref{Wcsi4}) with $\sigma={\sf a}_{ij}={\sf s}_{ij}=0$.

842: Special cases:

843: \begin{enumerate}

844: \item{} If $\widetilde{W}^{(0)}_{\hat{i}}=\Psi_{,\hat{i}}, ~ \widetilde{W}^{(0)}_i=\widetilde{H}^{(1)}=0$ then $(R)_{-1}=0$.

845: \item{} If $\widetilde{W}^{(0)}_{\hat{i}}=\widetilde{H}^{(1)}=\widetilde{H}^{(0)}=0$ then $(R)_{-1}=(R)_{-2}=0$.

846: \end{enumerate}

847: 

848: 

849: The connection $\nabla$ for the above metric can be decomposed as

850: \beq

851: \nabla=\widetilde{\nabla}-{\mbold{\tau}},

852: \eeq

853: where ${\mbold\tau}$ is a $(2,1)$ tensor (or operator acting in the obvious way) with only non-positive boost-weight components, while $\widetilde{\nabla}$ has the following property: {\sl There exists a null frame  such that the connection coefficients of $\widetilde{\nabla}$ has the following properties: }

854: \begin{enumerate}

855: \item{} Positive boost weight connection coefficients are all zero.

856: \item{} Zero boost weight connection coefficients are either all constants or are connection coefficients of $\mathcal{M}_H$.

857: \end{enumerate}

858: Henceforth we will assume that this null frame is chosen. Moreover, ${\mbold\tau}$ and $\widetilde{\nabla}$ can be chosen such that 

859: \beq

860: &&{\mbold\tau}_{{\bf e}_{\alpha}}X={\mbold\tau}_{X}{\bf e}_{\alpha}=0,\quad \nonumber \\

861: && \widetilde{\nabla}_{X}g=0, \quad g({\mbold\tau}_XY,Z)+g(Y,{\mbold\tau}_XZ)=0, \quad \forall X,Y,Z\in TM.

862: \eeq

863: Note that the torsion of $\widetilde{\nabla}_{X}$ is $S_XY={\mbold\tau}_YX-{\mbold\tau}_XY$. 

864: 

865: Let $R$ and $\widetilde{R}$ be the curvature tensors to the connections $\nabla$ and $\widetilde{\nabla}$ respectively. Then

866: \beq

867: R=\widehat{R}+\widetilde{R},

868: \eeq

869: where $(\widehat{R})_b=0$ for $b\geq 0$  while $(\widetilde{R})_b=0$ for $b>0$. Moreover, all boost weight 0 components are constants\footnote{This can be seen by explicit calculation of the curvature tensors.}. We also have

870: \begin{thm}

871: For all $k$, $\nabla^{(k)}{R}$ has maximal boost order 0 where all boost weight 0 components are constants.

872: \end{thm}

873: \begin{proof}

874: We will prove the Theorem by induction. Let us in the following denote the projection of a tensor $T$ onto the vector space spanned by the components of boost weight $b$ by $(T)_b$. The key observation is that the $(R)_0=(\widetilde{R})_0$ is of constant curvature over the $VSI$ fibers. As a result, ${\mbold\tau}({R})_0=0$. Thus the theorem is true for ${\nabla}({R})_0$. Moreover, by applying the Bianchi identity for ${\nabla}{R}$, we can show that it is true for ${\nabla}({R})_{-1}$. Hence, the theorem is true for $k=1$.

875: 

876: Assume the theorem is true for $k$. Then consider $\nabla\nabla^{(k)}R$. Immediately, we have that it is true for $\widetilde{\nabla}({\nabla}^{(k)}{R})_0$. Moreover, the curvature tensor is of constant curvature over the $VSI$ fibers, so ${\mbold\tau}({\nabla}^{(k)}{R})_0=0$; hence, the theorem is true for ${\nabla}({\nabla}^{(k)}{R})_0$

877: 

878: For ${\nabla}({\nabla}^{(k)}{R})_{-1}$ the critical part is the one that raises the boost weight by one. However, by using the identity

879: \beq

880: \left[{\nabla}_{A},{\nabla}_{B}\right]T_{C_1\dots C_i\dots C_p}=\sum_{i=1}^p{R}^{E}_{~C_i AB}T_{C_1\cdots E\cdots C_p},

881: \label{commutator}\eeq

882: recursively, and also the Bianchi identity, we see that it is also true 

883: for ${\nabla}({\nabla}^{(k)}{R})_{-1}$. Hence, by induction, the Theorem follows.

884: \end{proof}

885: These results consequently imply that the 'fibered {\sc vsi}' 

886: spacetime given by eq.(\ref{eq:csi}) is a $CSI$ spacetime. A crucial part of the proof is 

887: that $(R)_0$ is of constant curvature over the $VSI$ fibers, which implies 

888: that ${\mbold\tau}({\nabla}^{(k)}{R})_0=0$. 

889: 

890: \begin{thm}

891: Consider the metric (\ref{eq:csi}). If for all $k$ $\nabla^{(k)}{R}$ has maximal 

892: boost order 0, where all boost weight 0 components are constants, then 

893: there exists a homoegenous space $(\widetilde{\mathcal{M}},\tilde{g})$ with the same 

894: curvature invariants as the metric (\ref{eq:csi}).

895: \end{thm}

896: \begin{proof}

897: The proof essentially follows along the same lines as above by noticing that the functions $\widetilde{W}_i$ and $\widetilde{H}$ do not contribute to the curvature invariants. The homogeneous space can be obtained by setting all of these to zero. 

898: \end{proof}

899: 

900: 

901: \section{Discussion}

902: 

903: 

904: We have obtained a number of general results.

905: We have studied warped product $CSI$ spacetimes (Theorem 3.1).

906: We have presented the canonical form for the Kundt $CSI$ metric

907: (Theorem 4.1), and studied higher dimensional  Kundt

908: spacetimes in some detail. In sections 3 \& 5 we have

909: constructed higher dimensional examples of

910: $CSI$ spacetimes that arise as warped products and that belong to the

911: Kundt class. In the appendices we shall present

912: all of the 3-dimensional $VSI$ metrics,

913: explicitly construct the metrics in

914: the 4-dimensional $CSI$ spacetimes, and we shall establish the

915: canonical higher dimensional $VSI$ metric.

916: 

917: 

918: 

919: Motivated by our results and examples we propose the following conjectures:

920: 

921: \begin{con}[$\FC$ conjecture] A spacetime is {$CSI$} iff there exists a null frame 

922: in which the Riemann tensor and its derivatives can be brought into one of the following forms: either

923: \begin{enumerate}

924: \item{} the Riemann tensor and its derivatives are constant, in which case we have a

925: locally homogeneous space, or

926: \item{} the Riemann tensor and its derivatives are of boost order zero with constant boost weight zero components at each

927: order.  This implies that Riemann tensor is of type II or less.

928: \end{enumerate}

929: \end{con}

930: Whereas the proof of the `if' direction is trivial, the `only if' part is significantly more difficult and would

931: require considering second order curvature invariants.  We also point out that

932: if we relax {$CSI$} to {$CSI$}$_{1}$ then in (1)

933: we must now include all 1-curvature homogeneous spacetimes as well, since these naturally

934: imply {$CSI$}$_{1}$ and are not

935: necessarily homogeneous spaces.

936: 

937: Assuming that the above conjecture is correct and that there exists such a preferred

938: null frame, then

939: 

940: 

941: \begin{con}[$\KC$ conjecture] If a spacetime is {$CSI$}, then

942: the spacetime is either locally homogeneous or belongs to the

943: higher dimensional Kundt $CSI$ class.

944: \end{con}

945: 

946: Finally, we have that

947: \begin{con}[$\RC$ conjecture]

948: If a spacetime is {$CSI$}, then it can be constructed from locally homogeneous

949: spaces and $VSI$ spacetimes.

950: \end{con}

951: This construction can be done by means of fibering, warping and tensor sums.

952: 

953: From the results above and these conjectures, it is plausible that for $CSI$

954: spacetimes that are not locally homogeneous, the Weyl type is

955: $II$, $III$, $N$ or $O$, and that all boost weight zero terms are

956: constant.

957: 

958: 

959: We intend to study the general validity of these conjectures in future work. The relationship

960: to curvature homogeneous spacetimes is addressed in \cite{Mp}.

961: Support for these conjectures in the 4D case is discussed in the

962: next section.

963: 

964: 

965: 

966: \section{Four--dimensional $CSI$}

967: 

968: Let us now consider the 4D $CSI$ spacetimes in more detail. We

969: assume that a $CSI$ spacetimes is {\em either} of Petrov

970: (P)-type I or Plebanski-Petrov (PP)-type I, {\em or} of P-type II

971: and PP-type II (branch A or B of Table 1,

972: respectively). In branch A, we explicitly assume that the

973: spacetime is not of P-type II and PP-type II, otherwise we would

974: be in branch B. It is plausible that in branch A the

975: spacetimes must be of P-type I \emph{and} PP-type I (but we have not

976: established this here). Since the spacetime is not of P-type II

977: and PP-type II, it is necessarily completely backsolvable (CB)

978: \cite{Carminati}. Since all of the zeroth order scalar curvature

979: invariants are constant, there exists a frame in which the

980: components of the Riemann tensor are all constant, and the

981: spacetime is curvature homogeneous $CH_0$ (i.e., $CSI_0 \equiv

982: CH_0$). If the

983: spacetime is also $CH_1$, then it is necessarily locally

984: homogeneous $H$ \cite{Mp}. Therefore, if it is not locally

985: homogeneous, it cannot be $CH_1$. Finally, by considering

986: differential scalar invariants, we can obtain information on the

987: spin coefficents in the $CSI$ spacetime. Although a comprehensive

988: analysis of the differential invariants is necessary, a

989: preliminary investigation indicates that $\tau, \rho$ and $\sigma$

990: must all be constant and are likely zero. Therefore, in branch A,

991: there are severe constraints on those spacetimes that are not

992: locally homogeneous, and it is plausible that there are no such

993: spacetimes.

994: 

995: In branch B of Table 1, we have that the spacetime is necessarily

996: of P-type II and PP-type II. It then follows from Theorem 7.1 below,

997: that all boost weight zero terms are necessarily constant

998: (i.e., the spacetime is $\FCSI{0}$). The spacetime is then

999: either CB, in which case all of the results in branch A apply, or

1000: they are NCB and a number of further conditions apply (these

1001: conditions are very severe \cite{Carminati}). In either case,

1002: there are a number of different classes characterized by their

1003: P-type and PP-type (in each case at least of type II), and in each

1004: class there are a number of further restrictions. By investigating

1005: the differential scalar invariants we then find conditions on the

1006: spin coefficients. By considering each class separately, we can

1007: then establish that $\tau = \rho = \sigma =0$, and that the

1008: spacetime is necessarily $ \KC$. All of these

1009: spacetimes are constructed in Appendix B.

1010: 

1011: \begin{table}

1012:   \includegraphics[width=13cm]{000csiTable.eps}

1013:   \caption{4D $CSI$. All terms are defined in the text. 

1014:   A dashed line indicates a conjectured result. Footnotes:

1015: (1) this indicates that the spacetime is at least of type $II$

1016: (i.e., of type $II$, $III$, $N$ or $0$), (2) spacetimes in this set are

1017: described in the text, (3)

1018: all of the 4D $\KC$ spacetimes are displayed

1019: explicitly in the text.}\label{Table}

1020: \end{table}

1021: 

1022: 

1023: The results are summarized in Table 1. In branch A of Table 1, we

1024: have established that if the $CSI$ spacetime is not locally

1025: homogeneous, then necessarily it is of P-type I or PP-type I (and

1026: does {\em not} belong to $\KC$, $\FC$, or

1027: $\RC$), it belongs to $CSI_0 \equiv CH_0$, but not

1028: $CH_1$ and there are a number of further constraints arising from

1029: the non-CB conditions and the differential constraints (i.e.,

1030: constraints on the spin coefficients). This exceptional set is

1031: very sparse, and mostly likely empty; but in order to demonstrate

1032: this the conjectures in 4D must be proven, utilising a number of

1033: 4D invariants.

1034: 

1035: We have explicitly constructed the 4D $\FC$,

1036: $\KC$, and $\RC$, and established the

1037: relationship between these $CSI$ subsets themselves and with $CSI

1038: \backslash H$, thereby lending support to the conjectures in the

1039: previous section. The metrics of these $CSI$ spacetimes are

1040: described in Appendix B. From this calculation of all of the

1041: members of the set $\FC \bigcap \KC$, we see

1042: that they all belong to $\RC$. That is, we have

1043: established the uniqueness of $\RC$ within this

1044: particular context. In particular, we have found all $CSI$ in

1045: branch B of Table 1.

1046: 

1047: 

1048: 

1049: Finally, let us  consider Petrov (P)-type II and Plebanski-Petrov (PP)-type II

1050: Lorentzian $CSI$ spacetimes. We shall show that if all zeroth order

1051: curvature invariants are constant then the boost

1052: weight zero curvature scalars, $\Psi_{2}$, $\Phi_{11}$ and $\Phi_{02}$ are

1053: constant.  For these algebraically special

1054: spacetimes the converse follows straightforwardly.

1055: 

1056: \begin{thm}

1057: If all zeroth order

1058: curvature invariants are constant in Petrov-type II and Plebanski-Petrov-type II

1059: Lorentzian $CSI$ spacetimes, then the boost

1060: weight zero curvature scalars, $\Psi_{2}$, $\Phi_{11}$ and $\Phi_{02}$ are

1061: constant.

1062: \end{thm}

1063: \begin{proof}

1064: 

1065: We make use of the following curvature invariants

1066: \begin{equation}

1067: \begin{array}{ccc}

1068: w_{1}=\Psi_{ABCD}\Psi^{ABCD}, &

1069: r_{1}=\Phi_{AB\dot{A}\dot{B}}\Phi^{AB\dot{A}\dot{B}}, &

1070: r_{2}=\Phi_{AB\dot{A}\dot{B}}\Phi^{B\quad\dot{B}}_{\quad

1071: C\quad\dot{C}}\Phi^{CA\dot{C}\dot{A}}.

1072: \end{array}

1073: \end{equation}

1074: It is worth noting that in general there is an additional independent

1075: degree 4 Ricci invariant $r_{3}$; however, in

1076: PP-types II or less there always exists a syzygy for $r_{3}$ in terms of

1077: $r_{1}$ and $r_{2}$ (or else it vanishes).  This

1078: is analogous to the well known syzygy for the Weyl invariants,

1079: $I^{3}=27J^{2}$, signifying an algebraically special

1080: spacetime.

1081: 

1082: Assuming first that the Weyl and Ricci canonical frames are aligned, then

1083: the non-vanishing curvature scalars are

1084: $\Psi_{2}$, $\Psi_{4}=1$, $\Phi_{11}$, $\Phi_{20}=\Phi_{02}$,

1085: $\Phi_{22}=1$ and we have $w_{1}=6\Psi_{2}^2$,

1086: $r_{1}=2\Phi_{02}^2+4\Phi_{11}^2$, $r_{2}=-6\Phi_{02}^2\Phi_{11}$. Clearly

1087: $w_{1}$ constant implies $\Psi_{2}$

1088: constant.  If $\Phi_{02}(2\Phi_{11}-\Phi_{02})(2\Phi_{11}+\Phi_{02})\neq0$

1089: then $\Phi_{02}$ and $\Phi_{11}$ can be

1090: expressed in terms of $r_{1}$ and $r_{2}$; therefore, constant $r_{1}$ and

1091: $r_{2}$ gives constant $\Phi_{11}$ and

1092: $\Phi_{02}$.  If $\Phi_{02}=0$ or $\Phi_{02}=\pm 2\Phi_{11}$ then $r_{1}$

1093: can be used to give the same result.  In this

1094: aligned case we find that constant $R$, $w_{1}$, $r_{1}$ and $r_{2}$

1095: implies that all curvature scalars are constant

1096: (including the boost weight 0 scalars); therefore, we have a curvature

1097: homogeneous spacetime.

1098: 

1099: In the non-aligned case we refer to Carminati and Zakhary \cite{Carminati} where it was shown that for

1100: this PP-type, complete backsolving (CB) of the

1101: Carminati-Zakhary (CZ) invariants can be achieved and all curvature scalars

1102: can be expressed in terms

1103: of zeroth order invariants; thus constant CZ

1104: invariants implies constant curvature scalars.  In the exceptional case in which

1105: $\Psi_{0}=\Psi_{1}=0$ in the Ricci canonical frame,

1106: complete backsolvability is not possible since $\Psi_{3}$ and $\Psi_{4}$

1107: cannot be determined from any zeroth order

1108: curvature invariant \cite{Carminati}.  However, this exceptional case occurs

1109: when the Weyl and Ricci canonical frames are

1110: aligned, for P-type II $\Psi_{3}=0$ and $\Psi_{4}=1$ and the remaining

1111: curvature scalars were shown above to be

1112: constant (this is a particular instance of a not completely backsolvable (NCB)

1113: case becoming completely backsolvable; see

1114: comment in Ref. 3 of \cite{Carminati}).

1115: \end{proof}

1116: 

1117: 

1118: 

1119: 

1120: 

1121: 

1122: We have consequently also shown that for P-type II and PP-type II spacetimes, if the zeroth order

1123: invariants are constant then all curvature scalars

1124: are constant. These results support the higher dimensional conjectures discussed

1125: in the previous section.

1126: 

1127: 

1128: \section*{Acknowledgements} We would like to thank R. Milson, V. Pravda and A. Pravdova 

1129: for helpful comments. This work was supported by NSERC (AC and NP) and the Killam

1130: Foundation (SH).

1131: \newpage

1132: 

1133: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

1134: % APPENDICES

1135: \appendix

1136: 

1137: \section{3-dimensional $VSI$ metrics}

1138: 

1139: By analogy with the four dimensional $VSI$ spacetimes, we begin by

1140: considering a real null frame $e_{i}=\{l,n,m\}$ such that the only

1141: non-vanishing inner products are $l^{a}n_{a}=1=-m^{a}m_{a}$.  In

1142: three dimensions, the Riemann tensor is equivalent to the Ricci

1143: tensor, therefore all zeroth order invariants vanish if

1144: 

1145: \begin{equation}

1146: R_{ab}=R_{22}l_{a}l_{b}+2R_{23}l_{(a}m_{b)}.

1147: \end{equation}

1148: 

1149: We identify $R_{22} \sim \phi_{22}$ and $R_{23} \sim \phi_{12} +

1150: \phi_{21}$ and we use the n-dimensional version of the $VSI$

1151: theorem to conclude that the analogues of $\kappa$, $\sigma$ and

1152: $\rho$ must vanish.  In three dimensions we find that that $\kappa

1153: \sim \gamma_{311}$, $\sigma=\rho\sim \gamma_{313}$, and $\tau \sim

1154: \gamma_{312}$, where $\gamma_{ijk}$ are the Ricci-rotation

1155: coefficients.  Therefore, we have that a three dimensional

1156: spacetime is $VSI$ if and only if $\gamma_{311}=\gamma_{313}=0$

1157: and has Segre type \{3\}, \{(21)\} or \{(111)\} with the

1158: possibility that the (non-)vanishing of $\gamma_{312}$ may lead to

1159: distinct subclasses.

1160: 

1161: The method we use to explicitly obtain the three dimensional $VSI$

1162: spacetimes involves writing the four dimensional $VSI$ spacetimes

1163: in terms of real coordinates $(u,v,x,y)$ then considering the

1164: resulting frame when either $x=const.$ or $y=const.$  We set

1165: $\zeta=(x+i y)/\sqrt{2}$ and $W=(W_1 + i W_2)/\sqrt{2}$

1166: throughout, and in every case the restriction to $y=const.$ yields

1167: nothing new, hence the relevant null frame, in coordinates

1168: $(u,v,x)$, is

1169: 

1170: \begin{eqnarray}

1171: l=\partial_{v}, & n=\partial_{u}-[H+\frac{1}{2}(W_{1}^{2}+W_{2}^{2})]\partial_{v}+W_{1}\partial_{x}, & m=\partial_{x}

1172: \end{eqnarray}

1173: 

1174: Upon restriction to $x=const.$ most of the four dimensional $VSI$

1175: spacetimes become either flat or a special case of a null frame

1176: given in the following tables. In addition, whenever

1177: $\gamma_{312}$ is nonzero then it is always equal to $-1/x$.

1178: 

1179: \begin{table}

1180: \begin{center}

1181: \begin{tabular}{cl}

1182: $\gamma_{312}$ & Constraints on $W_{1}(u,v,x)$, $W_{2}(u,v,x)$ and $H(u,v,x)$ \\ \hline \\

1183: 

1184: $=0$ & A1) (P-III ; $\tau = 0$ ; PP-N)\\

1185:  & $W_{1}=W_{01}(u,x)$, $W_{2}=W_{02}(u,x)$, $H=vh_{1}(u,x)+h_{0}(u,x)$  \\

1186: \\

1187: $\neq 0$ & B1) (P-III ; $\tau \neq 0$ ; PP-N)\\

1188:  & $W_{1}=-\frac{2v}{x}+W_{01}(u,x)$, $W_{2}=W_{02}(u,x)$, $H=-\frac{v^2}{2x^2}+vh_{1}(u,x)+h_{0}(u,x)$

1189: 

1190: \end{tabular}

1191: \end{center}

1192: \caption{Segre type \{3\} i.e. $R_{22} \neq 0$ and $R_{23} \neq 0$.}

1193: \end{table}

1194: 

1195: \begin{table}

1196: \begin{center}

1197: \begin{tabular}{cl}

1198: $\gamma_{312}$ & Constraints on $W_{1}(u,v,x)$, $W_{2}(u,v,x)$ and $H(u,v,x)$ \\ \hline \\

1199: 

1200: $=0$ & D1) (P-N ; $\tau = 0$ ; PP-O)  special case of A1) \\

1201:  & $W_{1}=xf_{1}(u)+f_{0}(u)$, $W_{2}=xg_{1}(u)+g_{0}(u)$, $H=h_{0}(u,x)$  \\

1202: \\

1203: $\neq 0$ & F1) (P-III ; $\tau \neq 0$ ; PP-O) special case of B1) \\

1204:  & $W_{1}=-\frac{2v}{x}+W_{01}(u)$, $W_{2}=W_{02}(u)$, $H=-\frac{v^2}{2x^2}+\frac{v}{x}W_{01}(u)+h_{0}(u,x)$

1205: 

1206: \end{tabular}

1207: \end{center}

1208: \caption{Segre type \{(21)\} i.e. $R_{22} \neq 0$ and $R_{23} = 0$.}

1209: \end{table}

1210: 

1211: \begin{table}

1212: \begin{center}

1213: \begin{tabular}{cl}

1214: $\gamma_{312}$ & Constraints on $W_{1}(u,v,x)$, $W_{2}(u,v,x)$ and $H(u,v,x)$ \\ \hline \\

1215: 

1216: $=0$ & I1) (P-III ; $\tau = 0$ ; vacuum)\\

1217:  & $W_{1}=W_{01}(u)$, $W_{2}=W_{02}(u)$, $H=xh_{01}(u)+h_{02}(u)$  \\

1218: \\

1219: $\neq 0$ & L1) (P-N ; $\tau \neq 0$ ; vacuum)\\

1220:  & $W_{1}=-\frac{2v}{x}$, $W_{2}=0$, $H=-\frac{v^2}{2x^2}+\sqrt{2}x[xh_{01}(u)+h_{02}(u)]$

1221: 

1222: \end{tabular}

1223: \end{center}

1224: \caption{Segre type \{(111)\} i.e. $R_{ij}=0$.}

1225: \end{table}

1226: 

1227: Tables 2--4 provide a list of the three-dimensional $VSI$ spacetimes. 

1228: Here, we have chosen to characterise the 3D cases according to Segre type. In this regard, the classification in \cite{HallCapocci} of three dimensional spacetimes is worth noting. In 4D the Segre types are related to the PP types \cite{exsol}. We have also indicated the corresponding class of four dimensional $VSI$

1229: spacetimes (in brackets) from which the 3D solutions were obtained. It is also worth noting that

1230: some of the spacetimes given here may be equivalent and so these

1231: tables may reduce further. For example, in Table 4 the vanishing

1232: of the Ricci tensor implies that all spacetimes listed there are

1233: flat; therefore, using the full 3 parameter Lorentz group,

1234: $SO(1,2)$, it should be possible to set $\gamma_{312}=0$ (i.e.,

1235: the flat metric should appear in this table). In Tables 2 and 3

1236: it is necessary to consider the frame freedom left after the Ricci tensor has

1237: been brought to its canonical form. This can then be used to

1238: determine if $\gamma_{312}$ can be made to vanish; alternatively,

1239: it may be an invariant of this left over frame freedom and the

1240: metrics would be inequivalent. If $\gamma_{312}$ can be set to

1241: zero, then it still remains to be shown if a coordinate

1242: transformation can be found relating the two metrics.

1243: 

1244: 

1245: 

1246: 

1247: 

1248: %\appendix

1249: 

1250: \section{4D $CSI$}

1251: \label{app:4D}

1252: 

1253: We know that (Lorentzian) homogeneous spaces and $VSI$ spacetimes are

1254: necessarily $CSI$ spacetimes. Let us display all 4D spacetimes

1255: that are $\FC$ and $\KC$. We shall find

1256: that all such spacetimes are necessarily in $\RC$.

1257: 

1258: The 4D spacetimes in $\FC  \bigcap \KC$ are

1259: as follows.

1260: 

1261: 

1262: \subsection{}\label{AppB1} $\mathcal{I}(\mathrm{AdS}_4)$

1263: \beq

1264: \d s^2&=&e^{-2py}\left[2\d u\left(\d v+\widetilde{H}\d u+\widetilde{W}_x\d x+\widetilde{W}_y\d y\right)+\d x^2\right] +\d y^2,

1265: \label{eq:4Dcsi}\eeq

1266: where

1267: \beq

1268: \widetilde{W}_x &=& v\widetilde{W}_x^{(1)}(u,x)+ \widetilde{W}_x^{(0)}(u,x,y),\nonumber  \\

1269: \widetilde{W}_y &=& \widetilde{W}_y^{(0)}(u,x,y),\nonumber \\

1270: \widetilde{H}&=&\frac{v^2}{8}(\widetilde{W}^{(1)}_x)^2+v\widetilde{H}^{(1)}(u,x,y)+\widetilde{H}^{(0)}(u,x,y).\nonumber

1271: \eeq and $\widetilde{W}^{(1)}_x$ satisfy the 3D {$VSI$}

1272: equations (see Appendix A).

1273: 

1274: Since $\widetilde{W}_x^{(0)}(u,x,y)$ depends on $y$, in addition

1275: to $u$ and $x$, this metric is fibered.  Due to the terms

1276: $e^{-2py}$ and $\d y^2$ in the metric, the spacetime is a warped product. By

1277: redefining $\widetilde{W}_x$ in the metric above by omitting the

1278: term $v\widetilde{W}_x^{(1)}(u,x)$ (the '$vW$' term) and in

1279: $\widetilde{H}$ omitting the term

1280: $\frac{v^2}{8}(\widetilde{W}^{(1)}_x)^2$ (the '$v^2H$' term), we

1281: can rewrite the metric as the tensor sum of the redefined metric

1282: (\ref{eq:4Dcsi}) and the metric

1283: \beq \d s^2{_+}&

1284: =&e^{-2py}\left[2\d u\left(\d v+

1285: \frac{v^2}{8}(\widetilde{W}^{(1)}_x)^2 \d u+

1286: v\widetilde{W}_x^{(1)}\d x \right)+\d x^2\right] +\d y^2,

1287: \label{eq:4Dcsisum} \eeq 

1288: Clearly the metric can be written in

1289: terms of the combined operations of fibering, warping and tensor summing,

1290: and hence belongs to $\RC$.

1291: 

1292: 

1293: 

1294: 

1295: 

1296: 

1297: 

1298: \subsection{} $\mathcal{I}({\sf Sol})$\beq

1299: \d s^2&=&e^{-2py}\left[2\d u\left(\d v+\widetilde{H}\d u+\widetilde{W}_x\d x+\widetilde{W}_y\d y\right)\right]+e^{-2qy}\d x^2 +\d y^2,

1300: \label{eq:4Dcsi3}\eeq

1301: where

1302: \beq

1303: \widetilde{W}_x &=& \widetilde{W}_x^{(0)}(u,x,y),\quad

1304: \widetilde{W}_y = \widetilde{W}_y^{(0)}(u,x,y),\nonumber \\

1305: \widetilde{H}&=&v\widetilde{H}^{(1)}(u,x,y)+\widetilde{H}^{(0)}(u,x,y).\nonumber

1306: \eeq This metric is fibered and warped.

1307: 

1308: 

1309: 

1310: \subsection{} $\mathcal{I}(M_1\times M_2)$:

1311: \beq

1312: \d s^2&=&2\d u\left(\d v+{H}\d u+{W}_i{\bf m}^i\right)+\delta_{ij}{\bf m}^i{\bf m}^j,

1313: \label{eq:4Dcsi4}\eeq

1314: where

1315: \beq

1316: {W}_i  &=& vW^{(1)}_i+\widetilde{W}_i^{(0)}(u,x,y),\nonumber \\

1317: H &= &

1318: v^2H^{(2)}+v\widetilde{H}^{(1)}(u,x,y)+\widetilde{H}^{(0)}(u,x,y).\nonumber

1319: \eeq and  $W^{(1)}_i$ and $H^{(2)}$ are given in the table below.

1320: This metric is fibered and warped, and due to the '$vW$' and

1321: '$v^2H$' terms, can be written as a tensor sum.

1322: 

1323: \begin{tabular}{|c||c|c|c|}\hline

1324: $\delta_{ij}{\bf m}^i{\bf m}^j$ & $\d x^2+\d y^2$ & $a^2(\d x^2+\sin^{2}x\d y^2)$ & $\d x^2+e^{-2qx}\d y^2$ \\

1325: \hline \hline

1326: $W^{(1)}_i{\bf m}^i$ & $\alpha \d x+\beta \d y$ & $0$& $\alpha \d x +\beta e^{-qx}\d y$\\

1327: $H^{(2)}$ & $\frac 18(4\sigma+\alpha^2+\beta^2)$ &$\frac \sigma{2}$ & $\frac 18(4\sigma+\alpha^2+\beta^2)$ \\

1328: \hline

1329: \end{tabular}

1330: 

1331: \subsection{} $\mathcal{I}(\mathrm{(A)dS}_3\times \mathbb{R})$:

1332: Metrics of the form (\ref{eq:csi2}), (\ref{eq:csi3}) and

1333: (\ref{eq:csi4}). These metrics are fibered, warped and tensor

1334: summed.

1335: 

1336: 

1337: 

1338: \subsection{}\label{AppB5}$\mathcal{I}(\mathrm{(A)dS}_4)$: Metrics of the form (\ref{eq:csi2}), (\ref{eq:csi3}) 

1339: and (\ref{eq:csi4}). In addition, for  $\mathcal{I}(\mathrm{dS}_4)$

1340: \beq \d

1341: s^2&=&\cos^2(\sqrt{\sigma}y)\cos^2(\sqrt{\sigma}x)\left[2\d

1342: u\left(\d v+\widetilde{H}\d u+\widetilde{W}_i{\bf

1343: m}^i\right)\right]\nonumber \\ && +\cos^2(\sqrt{\sigma}y)\d x^2+\d

1344: y, \eeq where \beq \widetilde{W}_i = \widetilde{W}_i^{(0)}(u,x,y),

1345: \quad

1346: \widetilde{H}=\frac{v^2}{2}\sigma+v\widetilde{H}^{(1)}(u,x,y)+\widetilde{H}^{(0)}(u,x,y).

1347: \eeq and in the case of $\mathcal{I}(\mathrm{AdS}_4)$, similar

1348: versions of the metrics (\ref{eq:csi3}) and (\ref{eq:csi4}) are obtained. These

1349: metrics are fibered, warped and tensor summed.

1350: 

1351: 

1352: 

1353: All of the metrics in sections \ref{AppB1}-\ref{AppB5} can be written in terms of the

1354: combined operations of fibering, warping and tensor summing.

1355: Therefore, all of the spacetimes constructed belong to

1356: $\RC$. We consequently have the result that in 4D

1357: $\FC$, $\KC$ and $\RC$ are

1358: equivalent. It is plausible that a similar result applies in

1359: higher dimensions.

1360: 

1361: 

1362: Finally, let us consider the tensor sum in a little more detail.

1363: Let us consider the $\RC$ metric (\ref{eq:csi2}).

1364: Redefining the metric by omitting the '$v^2H$' term, it can be

1365: rewritten as the 'tensor sum' of the redefined metric and

1366: \beq 

1367: \d s^2{_+}&=& \left[2\d u\left(\d

1368: v+\frac{v^2}{2}\sigma \d u + 2\sqrt{\sigma}\tan(\sqrt\sigma x) \d

1369: x\right)\right] +\d x^2. \label{eq:csi2sum}\eeq The term

1370: $\left[2\d u\left(\d v+\frac{v^2}{2}\sigma \d u\right)\right]$ can

1371: be rewritten as $\left[\left(1+ \frac{\sigma}{2}UV\right)^{-2} \d U \d

1372: V\right]$, which is a metric of constant Ricci scalar curvature.

1373: 

1374: It is plausible that we have constructed all of the 4D $CSI$ spacetimes here.

1375: To prove this, we must prove the conjectures outlined above. This is perhaps best

1376: done by considering higher order differential scalar

1377: invariants in 4D \cite{4DVSI}, which we hope to do in a future

1378: paper.

1379: 

1380: 

1381: \section{VSI metrics}

1382: We can now  write down the metric

1383: for higher dimensional VSI spacetimes in a canonical form. An immediate corollary of Theorem \ref{thm:Kundt} is:

1384: \begin{cor}

1385: Any {$VSI$} metric can be written

1386: \beq

1387: \d s^2=2\d u\left[\d

1388: v+H(v,u,x^k)\d u+W_{i}(v,u,x^k)\d x^i\right]+\delta_{ij}\d x^i\d x^j.

1389: \eeq

1390: \end{cor}

1391: \begin{proof}

1392: In \cite{Higher} it was proven that all {$VSI$} metrics have $L_{ij}=0$ and $R_{\hat{i}\hat{j}\hat{k}\hat{l}}=0$. The first condition implies that the metric is Kundt, hence of the form (\ref{Kundt}), while the second implies that the spatial metric, $\tilde{g}_{ij}$, is flat. Using Theorem \ref{thm:Kundt} the corollary now follows.

1393: \end{proof}

1394: 

1395: We are now in a position to determine the explicit metric forms

1396: for higher dimensional $VSI$ spacetimes \cite{CMPPPZ}, which we

1397: shall return to in future work.

1398: 

1399: \newpage

1400: 

1401: 

1402: \begin{thebibliography}{99}

1403: \bibitem{PTV}

1404: F. Pr\"ufer, F. Tricerri, L. Vanhecke, 1996, Trans. Am. Math. Soc.

1405: \textbf{348},  4643.

1406: 

1407: 

1408: \bibitem{4DVSI}

1409: V. Pravda, A. Pravdov\' a, A. Coley and R. Milson,,

1410: 2002, Class. Q. Grav. {\bf 19}, 6213 [gr-qc/0209024].

1411: 

1412: \bibitem{BijPod} J. Bicak and J. Podolsky, 1999, J. Math. Phys.

1413: {\bf 40}, 4495;  J. Podolsky and M. Ortaggio, 2003, Class.

1414: Q. Grav. {\bf 20}, 1685 \& J. B. Griffiths, P. Docherty and J.

1415: Podolsky, 2004, Class. Q. Grav. {\bf 21}, 207.

1416: 

1417: 

1418: \bibitem{obuk} Y. N. Obukhov, 2004, Phys. Rev. D {\bf 69}, 024013.

1419: 

1420: 

1421: \bibitem{Siklos} S.T.C. Siklos, 1985,

1422: {\it Lobatchevski plane wave gravitational waves, in Galaxies,

1423: Axisymmetric systems and Relativity}, ed. M. MacCallum (Cambridge

1424: University Press, Cambridge).

1425: 

1426: \bibitem{HJS} H. -J. Schmidt, 1998, Int. J. Th. Phys. {\bf 37}, 1691.

1427: 

1428: \bibitem{Hervik} S. Hervik, 2004, Class. Q. Grav. {\bf 21}, 4273.

1429: 

1430: 

1431: \bibitem{epsilon} N. Pelavas, A. Coley, R. Milson, V. Pravda and A. Pravdova, 2005, J. Math. Phys. \textbf{46} 063501

1432: [gr-qc/0503040].

1433: 

1434: \bibitem{class}  A. Coley, R. Milson, V. Pravda and A. Pravdova,

1435: 2004, Class. Q. Grav. {\bf 21}, L35 [gr-qc/0401008].

1436: 

1437: 

1438: \bibitem{Milson} R. Milson, A. Coley, V. Pravda and A. Pravdova, 2005, 

1439: Int. J. Geom. Meth. Mod. Phys. {\bf 2},

1440: 41 [gr-qc/0401010].

1441: 

1442: 

1443: \bibitem{Carminati} J. Carminati and E. Zakhary, 2002,  J. Math. Phys. {\bf 43}, 4020.

1444: 

1445: 

1446: \bibitem{Hall} G.S.  Hall, 2003, {\em Symmetries and Curvature Structure

1447: in General Relativity} (World Scientific, Singapore).

1448: 

1449: 

1450: 

1451: 

1452: \bibitem{Mp} R. Milson and N. Pelavas, 2006; preprint.

1453: 

1454: 

1455: 

1456: \bibitem{Singer} I. M. Singer, 1960, Comm. Pure and App. Math. {\bf 13}, 685.

1457: 

1458: \bibitem{Higher}    A. Coley, R. Milson, V. Pravda and A. Pravdova, 2004,

1459: Class. Q. Grav. {\bf 21}, 5519.

1460: 

1461: \bibitem{PS} F. Podesta and A. Spiro, 1995, preprint {\em Introduzione ai gruppi

1462: transformation}.

1463: 

1464: 

1465: 

1466: \bibitem{exsol} H. Stephani,D. Kramer,M. MacCallum and E. Herlt, 2003,

1467: \textit{Exact Solutions to Einstein's Field Equations},

1468: (Second Edition, Cambridge Univ. Press).

1469: 

1470: 

1471: 

1472: \bibitem{Petrov} A. Z. Petrov, 1969, \textit{Einstein Spaces} (Pergamon Press).

1473: 

1474: \bibitem{bianchi} V. Pravda, A. Pravdov\' a, A. Coley and R. Milson, 2004,

1475: Class. Quant. Grav. {\bf 21},  2873 [gr-qc/0401013].

1476: 

1477: \bibitem{Milnor} 

1478: J. Milnor, 1976, \textit{Adv. Math.} \textbf{21} 293

1479: 

1480: \bibitem{CMPPPZ}    A. Coley, R. Milson, V. Pravda, A. Pravdova, N.

1481: Pelavas and R. Zaleletdinov 2003,  Phys. Rev. D. 

1482: {\bf 67}, 104020 [gr-qc/0212063].

1483: 

1484: \bibitem{HallCapocci}

1485: G.S. Hall and S. Capocci, 1999, J. Math. Phys. \textbf{40}, 1466

1486: 

1487: \end{thebibliography}

1488: \end{document}

1489: