0707.0279/ms.tex
1: %\documentclass[12pt,preprint]{aastex}
2: \documentclass[pra,a4paper,preprintnumbers,aps,superscriptaddress,twocolumn,floatfix,showpacs,showkeys]{emulateapj}
3: %
4: \usepackage{latexsym}
5: \usepackage{epsfig}
6: \usepackage{amssymb}
7: %
8: %\bibliographystyle{authordate1}
9: %\bibliographystyle{aps}
10: %
11: \def\bg#1{\mbox{\boldmath$#1$}}
12: %
13: \newcommand{\Tr}{\mbox{Tr}}
14: \newcommand{\tr}{\mbox{tr}}
15: \newcommand{\Det}{\mbox{det}}
16: \newcommand{\Dim}[1]{\mbox{dim[$#1$]}}
17: \newcommand{\del}{\partial}
18: \newcommand{\bea}{\begin{eqnarray}}
19: \newcommand{\eea}{\end{eqnarray}}
20: \newcommand{\be}{\begin{equation}}
21: \newcommand{\ee}{\end{equation}}
22: \newcommand{\bk}{{\bf k}}
23: \newcommand{\bp}{{\bf p}}
24: \newcommand{\bq}{{\bf q}}
25: \newcommand{\br}{{\bf r}}
26: \newcommand{\bx}{{\bf x}}
27: \newcommand{\om}{\omega}
28: \newcommand{\Om}{\Omega}
29: \newcommand{\ua}{\uparrow}
30: \newcommand{\da}{\downarrow}
31: %\newcommand{\la}{\leftarrow}
32: \newcommand{\ra}{\rightarrow}
33: \newcommand{\e}{\epsilon}
34: \newcommand{\ve}{\varepsilon}
35: \newcommand{\nn}{\nonumber}
36: \newcommand{\ket}[1]{\mbox{$\mid\!#1\rangle$}}
37: \newcommand{\bra}[1]{\mbox{$\langle#1\!\mid$}}
38: \newcommand{\wh}[1]{\widehat{#1}}
39: \newcommand{\ex}[1]{\langle\,#1\,\rangle}
40: \newcommand{\D}{{\cal D}}
41: \newcommand{\lam}{\bar{\lambda}}
42: \newcommand{\Mvariable}{}
43: \newcommand{\half}{{1\over 2}}
44: %
45: \begin{document}
46: %
47: \title{Accelerating Cosmologies with an Anisotropic Equation of State}
48: % or
49: %\title{Cosmology with an Anisotropic Equation of State and the CMB Anomalies}
50: % or
51: %\title{Dark Energy with an Anisotropic Equation of State}
52: % or
53: %\title{Cosmology with an Anisotropic Equation of State}
54: %\date{\today}
55: \author{Tomi Koivisto}
56: %\email{tomikoiv@pcu.helsinki.fi}
57: \affil{Helsinki Institute of Physics, P.O. Box 64, FIN-00014 Helsinki, Finland}
58: %\affil{Department of Physical Sciences, Helsinki University, P.O. Box 64, FIN-00014 Helsinki, Finland}
59: \author{David F. Mota}
60: %\email{D.Mota@thphys.uni-heidelberg.de}
61: \affil{Institute for Theoretical Physics, University of Heidelberg, 69120 Heidelberg,Germany}
62: %
63: \begin{abstract}
64: If the dark energy equation of state is anisotropic, the expansion rate of the 
65: universe becomes direction-dependent at late times. We show that such models are not only cosmologically viable but that 
66: they could explain some of the observed anomalies in the CMB. The possible anisotropy can then be constrained by studying its effects on the 
67: luminosity distance-redshift relation inferred from several observations. A vector field action for dark energy is also presented as an example of 
68: such possibility.
69: \end{abstract}
70: %
71: \keywords{(cosmology:) cosmic microwave background;
72: cosmology: miscellaneous;
73: cosmology: observations;
74: cosmology: theory;
75: cosmology: large-scale structure of universe}
76: \maketitle
77: %
78: \section{Introduction}
79: %
80: It is not clear whether the large-angle anomalies in the observed cosmic microwave 
81: background are of a cosmological origin and not due to systematics \citep{Eriksen:2003db,Land:2005ad,Copi:2005ff}. 
82: Nevertheless, it is tempting to associate the apparent statistical anisotropy with dark energy,
83: since the anomalies occur at the largest scales, and these enter inside the horizon
84: at the same epoch that the dark energy dominance begins. 
85: 
86: The paramount characteristic of dark energy is its negative pressure. We examine then the possibility that this pressure varies with the direction.
87: Then also the universal acceleration becomes anisotropic, and one would indeed see otherwise unexpected effects at 
88: the smallest multipoles of CMB. 
89: At the background level we will find a similar CMB pattern as in an universe which is ellipsoidal at the era of last 
90: scattering\citep{Campanelli:2006vb}. We then show that the supernovae could be used to distinguish these different 
91: scenarios and to constrain the possible anisotropic properties of dark energy. 
92: 
93: There are several motivations for anisotropic models of dark energy. For
94: instance,  many dark energy models are in principle 
95: compatible with the FLRW metric but exhibit anisotropic stresses at the perturbative level,
96: including nonminimally coupled fields, viscous fluids and modified gravity models
97: \citep{Koivisto:2005mm,uzan,mota,tomi1,tomi2}. It is possible that a more accurate
98: description of such models should take into account the anisotropic effects on the background expansion, 
99: which then breaks the statistical isotropy of perturbations too (as appears to have
100: happened in the CMB \citep{Eriksen:2003db,Land:2005ad,Copi:2005ff}).  
101: Also, considering dark energy as an effective description of a backreaction of sizable 
102: inhomogeneities in dark matter \citep{backreaction,back}, the validity of a perfect fluid description could seem 
103: dubious. 
104: 
105: To have a general description of an anisotropic dark energy component, we consider a phenomenological parameterization of dark energy
106: in terms of its equation of state ($w$) and two skewness parameters ($\delta$, $\gamma$) and include also a 
107: coupling term ($Q$) between dark energy and a perfect fluid (dark matter). 
108: Previous studies anisotropic dark energy have mainly considered the anisotropic properties of the inhomogeneous perturbations 
109: \citep{Koivisto:2005mm, Battye:2006mb} 
110: whereas our approach here is to focus on a smooth cosmology with the anisotropic pressure field. Whereas in a
111: previous approach the anisotropy was only weakly constrained \citep{mota}, we find that in the present 
112: description only a narrow parameter range can survive the observational 
113: tests. 
114: Similar anisotropic inflation has been considered in the early universe \citep{Burd:1991ew,Ford:1989me}. 
115: Therefore we focus on the present paper more on the dark energy era, where the qualitative differences are that 
116: matter cannot be neglected and that the cosmologies do not isotropize.
117: As a proof of a concept, we write also an explicit field theory example where a vector field drives the anisotropic 
118: acceleration of the universe. 
119: 
120: An anisotropic expansion is not compatible with the 
121: Robertson-Walker (RW) metric. Hence we use the B(I) (Bianchi type I)  metric which 
122: generalizes the flat RW metric and may be employed 
123: to obtain limits on cosmological skew pressures from the CMB\citep{Barrow:1997sy}.
124: The line-element of a B(I) universe is:   
125: \be \label{metric}
126: ds^2 = -dt^2 +a^2(t)dx^2+b^2(t)dy^2 + c^2(t)dz^2.
127: \ee
128: There are thus three scale factors, and consequently three expansion rates. In principle all these
129: could be different, and in the limiting case that all of them are equal one  recovers the
130: RW case. It is useful to express the mean expansion rate as an average Hubble rate $H$
131: \be
132: H \equiv \frac{1}{3}\left(\frac{\dot{a}}{a}+\frac{\dot{b}}{b} + \frac{\dot{c}}{c}\right),
133: \ee
134: (where an overdot means derivative wrt $t$) and then the differences of the expansion rates 
135: as the Hubble-normalized shear $R$ and $S$ \citep{Barrow:1997sy}
136: \be
137: R  \equiv \frac{1}{H}\left(\frac{\dot{a}}{a}-\frac{\dot{b}}{b}\right), \quad
138: S  \equiv \frac{1}{H}\left(\frac{\dot{a}}{a}-\frac{\dot{c}}{c}\right).
139: \ee
140: We consider a universe filled with a perfect fluid having the energy-momentum tensor 
141: $
142: T^\mu_{(m)\nu} = diag(-1,w_m,w_m,w_m)\rho_m,
143: $
144: and dark energy, which we allow to have the most general energy-momentum tensor compatible with 
145: the metric (\ref{metric})
146: \be
147: T^\mu_{\phantom{\mu}\nu} = diag(-1,w,w+3\delta,w + 3\gamma)\rho.
148: \ee
149: The generalized Friedmann equation may be written as
150: \be
151: H^2 = \frac{8\pi G}{3}\frac{\rho_m+\rho}{1-\frac{1}{9}\left(R^2+ S^2 - RS\right)}.
152: \ee
153: We let the two components also interact. The continuity equations are then
154: \be
155: \dot{\rho}_m + 3H\left(1+w_m\right)\rho_m = Q H \rho,
156: \ee
157: and 
158: \be
159: \dot{\rho}   + 3\left[\left(1+w\right)H + \delta \frac{\dot{b}}{b} + 
160: \gamma\frac{\dot{c}}{c} \right]\rho = -Q H \rho,
161: \ee
162: where $Q$ determines the coupling. If it vanishes, it follows that $\rho_m \sim (abc)^{-1-w_m}$ and $\rho \sim 
163: (abc)^{-1-w}b^{3\delta}c^{3\gamma}$. 
164: Defining $x \equiv \frac{1}{3}\log{(-g)}$, where the metric determinant $g=-abc$,
165: one notes that $H = \dot{x}$. We will use $x$ as our time variable rather than $t$. Derivative wrt $x$ is
166: denoted by star. We also define the dimensionless density fractions
167: $$
168: \Omega_m \equiv \frac{8\pi G}{3}\frac{\rho_m}{H^2}, \qquad U \equiv
169: \frac{\rho}{\rho_m+\rho}.
170: $$ 
171: Using $R$, $S$, and $U$ as our dynamical variables, the system can finally be written as
172: %
173: \bea \label{system}
174:   U^*&=&   U\left(U -1\right)     \left[ \gamma \left(3+R-2S \right)  +\right.\nonumber\\&&\left.
175: \delta\left(3-2R+S\right) + 3\left(w - w_m\right) \right] - UQ\nonumber,
176: \\
177:   S^* & = &
178: \frac{1}{6}\left(9 - R^2 + RS - S^2 \right) \\ \nonumber
179: & x & \left\{ S\left[U\left(\delta + \gamma+w-w_m\right)+w_m-1\right]-6\gamma U \right\},
180: \\ \nonumber
181:   R^* & = &
182: \frac{1}{6}\left(9 - R^2 + RS - S^2 \right) \\ \nonumber
183: & x & \left\{ R\left[U\left(\delta + \gamma+w-w_m\right)+w_m-1\right]-6\delta U \right\}.
184: \eea
185: %
186: Note the coupling term $Q$ appears only in the evolution 
187: equation for $U$. Nevertheless its presence can change the dynamics completely \citep{us}. 
188: 
189: \section{Vector Field}
190: 
191: As a proof of concept, we present an explicit field theory for the anisotropically stressed dark energy. 
192: Consider the vector field action
193: \be\label{vector}
194: S = \int d^4 x \sqrt{-g} \left[\frac{R}{16\pi G} -\frac{1}{4}F_{\mu\nu}F^{\mu\nu} - V + qL_m\right]
195: \ee
196: where the kinetic term involves $F_{\mu\nu} = \partial_\mu A_\nu - \partial_\nu A_\mu$.
197: The potential $V$ and the possible couplings $q$ are understood as functions of the quadratic 
198: term $A^2 = A_\mu A^\mu$. The field equations are
199: $
200: G_{\mu\nu} = 8\pi G\left( T^m_{\mu\nu} + T^A_{\mu\nu}\right),
201: $
202: where the energy-momentum tensor of the vector field follows by varying wrt the metric
203: \be \label{emt}
204: T^A_{\mu\nu} = F_{\mu\alpha}F_{\nu}^\alpha + 2V' A_\mu A_\nu - \left( \frac{1}{4} F_{\alpha\beta} 
205: F^{\alpha\beta} + V \right) g_{\mu\nu}.
206: \ee
207: Prime denotes derivative wrt $A^2$.
208: With the metric (\ref{metric}), the equation of motion for the $A_0$ component of the field 
209: dictates that $[q' L_m + V']A_0=0$. Thus we confine to purely spatial vector 
210: fields. The spatial components of the energy-momentum tensor are,
211: $$
212: T^{iA}_{\phantom{i}j} = \frac{1}{a_i^2}\left( -\dot{A}_i\dot{A}_j + 2V' A_iA_j\right) +
213:         \left(\frac{1}{2}\sum_{k=1}^3 \frac{A_k^2}{a_k^2} - V\right)\delta^i_j.
214: $$
215: The off-diagonal terms should vanish. 
216: Thus one is restricted to consider only vectors
217: which are parallel to one of the coordinate axis. There, however, could be many such fields.
218: 
219: In the RW case all the diagonal components  of $T^i_{\phantom{i}j}$ should be equal, which
220: indeed requires three vector fields, one in each coordinate direction, of exactly equal
221: magnitudes, as the triads of Armendariz-Picon\citep{Armendariz-Picon:2004pm}. With the 
222: metric (\ref{metric}), the only constraint is that every vector field should be
223: along a coordinate axis. Any system of these vector fields is thus {\it required} to have 
224: an anisotropic equation of state, unless it reduces to the very special triad case. Let us therefore briefly
225: look at the case of a single vector field with $A_1=A_3=0$. 
226: Notice that, in analogy to system (\ref{system}), the coupling $Q$ is now related to the function $q$. If we 
227: define the additional dimensionless parameters 
228: $$
229: X \equiv \frac{\dot{A}_2}{b H}, \qquad Y \equiv \frac{V}{H^2}, \qquad Y_1 = 2\frac{V'A_2}{b^2 H^2}, 
230: $$
231: one can relate the present case to the general notation by noting that
232: $H^2\rho = X^2/2 + Y$, and that if $S=0$, then  
233: \be
234: w = \frac{X^2-2Y}{X^2+2Y}, \quad 
235: \delta = \frac{-X^2-Y_1}{\frac{3}{2}X^2 + 3Y}. 
236: \ee
237: We see that the anisotropy is naturally small if the field is either subdominant or near its minimum.
238: Such models, even though perturbatively close to standard cosmology, would be excluded by imposing the RW symmetry.
239: Viable models with large anisotropy do also exist as shown below. An extensive investigation of specific models will 
240: be undertaken elsewhere \citep{us}.
241: 
242: \section{Scaling solutions (in the axisymmetric case)}
243: 
244: %
245: 
246: To begin, we will assume for simplicity that 1) the perfect fluid is 
247: minimally coupled dark matter, $w_m=Q=0$, and further that 2) the skewness together
248: with the equation of state of dark energy is constant, $\dot{w}=\dot{\delta}=\dot{\gamma}=0$ (see \citep{us} for an extensive analysis with these assumptions relaxed).  
249: In our model the universe is then initially (close to) isotropic, but is driven to anisotropic expansion
250: by the skewness of dark energy.  Under the additional simplifying assumption
251: of axial symmetry ($S=\gamma=0$), in FIG. \ref{outcomes} we indicate the final possible stages of such universe, corresponding to the fixed 
252: points of the system (\ref{system}).
253: %
254: \begin{figure}[ht]
255: \begin{center}
256: \includegraphics[width=0.4\textwidth,height=0.4\textwidth]{f1.eps}
257: \caption{\label{outcomes} Asymptotic state  from an Einstein-deSitter stage.
258: The future fate depends on the dark energy properties $w$ and $\delta$ and is classified into three
259: possibilities. If $\delta+w > 0$, the isotropically expanding dust 
260: domination (solution (\ref{flrw_s}) for which $U=0$) 
261: continues forever. Otherwise, the universe will end up expanding anisotropically 
262: and either dominated by dark energy (solution (\ref{domi_s}) for which $U=1$) or 
263: exhibiting a scaling property (solution (\ref{scal_s}) for which $0<U<1$).}
264: \end{center}
265: \end{figure}
266: %
267: The universe will then end up in three possible scenarios: (I) The isotropic RW case corresponding to 
268: \be \label{flrw_s}
269: R=0, \quad U = 0.
270: \ee
271: (II) An anisotropic dark energy dominated solution with
272: \be \label{domi_s}
273: R = \frac{6\delta }{w+\delta-1}, \quad U =1.
274: \ee 
275: Or (III) a scaling solution,
276: \be \label{scal_s}
277: R = \frac{3\left(w+\delta\right)}{2\delta}, \quad
278: U = 
279: \frac{-\left(w+\delta\right)}{3\delta^2-2\delta w - w^2}.
280: \ee
281: Given that $w$ is negative, this solution accelerates
282: $w_{eff} \equiv -\frac{2}{3}\frac{H^*}{H}-1<-\frac{1}{3}$, if
283: $w<\delta<\frac{w}{3}$ or $-\frac{w}{5}<\delta<-\frac{w}{3}$. 
284: Note that in the $R=0$ case scaling
285: solutions could only be found for coupled components.
286: 
287: Notice that within the RW universe it has been proven difficult to address the coincidence problem by finding a model 
288: entering from a matter-dominated scaling solution to an accelerating scaling solution.
289: Allowing for the presence of three expansion rates opens up the possibility of  
290: describing a universe entering from a perfect fluid dominated scaling to 
291: an anisotropically accelerating scaling era. This might eventually help to understand the 
292: coincidence problem, since then matter and dark energy would have had similar energy densities
293: both in the past and in the future (in the past the dark energy fraction, if constant 
294: should not exceed about $1/10$ \cite{georg}).
295: 
296: 
297: We now study the observational implications of models with nonzero 
298: skewness, relaxing the assumption 2). The relevant case is then a universe entering from RW solution 
299: (\ref{flrw_s}) to an anisotropically accelerating universe (which generalizes (\ref{domi_s}) 
300: if $\gamma \neq 0$).
301: 
302: \section{CMB anisotropy}
303: 
304: The B(I) background predicts a quadrupole pattern in the CMB. 
305: It is not clear how the fluctuations in the photon temperature
306: distribution evolve when the background expansion becomes anisotropic, 
307: since the anisotropic sources, even if perturbatively small, couple to the 
308: perturbations at the first order \citep{us}.
309: It seems inevitable that statistically anisotropic features would then be, in principle,
310: present in the whole spectrum of fluctuations (though in our case such features
311: would be confined to larger scales as the anisotropy of the expansion becomes
312: important only recently). Several large-angle anomalies have been reported
313: in the CMB \citep{Eriksen:2003db,Land:2005ad,Copi:2005ff}. It has been demonstrated
314: that CMB spectrum of the tilted B(VIIh) background \citep{Hawking:1968zw,1985MNRAS.213..917B,Pontzen:2007ii}
315: would have potential to model these anomalies \citep{Jaffe:2005pw} were it not in conflict with a dark energy 
316: dominated universe at the smaller angular scales \citep{Jaffe:2005gu}. 
317: 
318: The B(I) model reproduces the predictions of the concordance model at small scales, while 
319: featuring anomalies at large angles, since the (originally statistically isotropic) CMB field 
320: experiences an (statistically) anisotropic integrated Sachs-Wolfe effect in the nowadays ellipsoidal 
321: universe. Furthermore, it is plausible that these anomalies can match the observed ones, since
322: a B(VIIh) model can be effectively described as a B(I) model with an additional anisotropic 
323: energy source\footnote{One may obtain the B(VIIO) model from B(I) by adding anisotropic curvature.
324: To obtain B(VIIh) model one should add also an isotropic curvature similar to such present
325: in B(V) models.} 
326: 
327: As a first step, we  check whether the background (\ref{metric}) is 
328: compatible with the available cosmological data. There are a number of calculations of  
329: CMB anisotropies in general Bianchi universes already available in the
330: literature \citep{Hawking:1968zw,1985MNRAS.213..917B,Pontzen:2007ii}. Here we consider the dark energy created quadropole and the 
331: effects to the luminosity distance.
332: 
333: By considering the geodesic equation for photons, one can derive an equation for the redshift $z$ of a 
334: photon arriving from the a direction $\hat{{\bf p}}$ 
335: \be 
336: \label{z_ellips}
337: 1+z(\hat{{\bf p}}) = 
338: %\left( \sum_i \frac{\hat{p}^2_i}{a_i^2}\right)^\frac{1}{2}=
339: \frac{1}{a}\sqrt{1 + \hat{p}_y^2e_y^2 + \hat{p}_z^2e_z^2}
340: \ee
341: in terms of the eccentricities
342: \be \label{ellips}
343: e_y^2 = \left(\frac{a}{b}\right)^2-1, \quad
344: e_z^2 = \left(\frac{a}{c}\right)^2-1.
345: \ee
346: Note that the scale factors and eccentrities here are evaluated at the time of 
347: last scattering in the case that the scale factors are all normalized to unity
348: today. If $T_*$ is the temperature at decoupling,
349: the temperature field is given by $T(\hat{{\bf p}}) = T_*/(1+z(\hat{{\bf p}}))$,
350: and its spatial average is $4\pi\bar{T} = \int d\Omega_{\hat{{\bf p}}}T(\hat{{\bf p}})$.
351: The anisotropy field is then
352: \be
353: \frac{\delta T(\hat{{\bf p}})}{\hat{T}} = 1- \frac{T(\hat{{\bf p}})}{\bar{T}}. 
354: \ee
355: The coefficients in the spherical expansion of this anisotropy field are called
356: $a_{\ell m}$, and due to orthogonality of spherical harmonics $Y_{\ell m}$ are given
357: by
358: \be
359: a_{\ell m} = \int d\Omega_{\bf p} \frac{\delta T(\hat{{\bf p}})}{\hat{T}} Y^*_{\ell m}.
360: \ee
361: The multipole spectrum is described by 
362: \be
363: Q_\ell = \sqrt{\frac{1}{2\pi}\frac{\ell(\ell+1)}{(2\ell+1)}\sum_{m=-\ell}^{\ell}|a_{\ell m}|^2}.
364: \ee
365: Expanding the redshifts (\ref{z_ellips}) in powers of the eccentricities (\ref{ellips}) one 
366: notes that the $a_{\ell m}$ will be real (since there is only even dependence on 
367: the polar angle in the anisotropy field and the imaginary parts of the $e^{i m\phi}$ integrate 
368: to zero), and that for all odd $\ell$ the $a_{\ell m}$ will vanish (since only even powers
369: of the azimuthal angle appear in the expansion). To first order in $e_{y,z}^2$, 
370: $\bar{T} = aT_*[1-\frac{1}{6}(e_z^2+e_y^2)]$, and in addition to the monopole, there is only 
371: the quadrupole
372: \bea \label{quad}
373: a_{20}&=&\frac{1}{3}\sqrt{\frac{\pi}{5}}\left(2e_z^2-e_y^2\right), \qquad
374: a_{21}=a_{2-1}= 0, \\ a_{22}&=&a_{2-2}= -\sqrt{\frac{\pi}{30}}e_y^2
375: \quad \Rightarrow \quad 
376: Q_2 = \frac{2}{5\sqrt{3}}\sqrt{e_z^4+e_y^4-e_z^2e_y^2}
377: \nonumber
378: \eea
379: 
380: The observed value of $Q_2$ is lower than the concordance model predicts \citep{Hinshaw:2006ia}. It has been suggested 
381: in previous works  that this discrepancy could be explained by an ellipsoidality of 
382: the universe \citep{Campanelli:2006vb,BeltranJimenez:2007ai}. This would require that the 
383: anisotropy of the background is suitably oriented with respect to the intrinsic quadrupole and cancels its power to 
384: a sufficient amount. Too large anisotropy would of course only make the situation worse regardless of the orientation.
385: Depending on the cosmological model, one should have $Q_2 \lesssim 2\cdot 10^{-5}$ to be consistent with observations 
386: taking into account the cosmic variance. The constraints this implies on the skewness of dark energy are very tight. 
387: However, we remark that in more general models, in particular with time-varying $\delta$ and $\gamma$, one could 
388: allow more
389: anisotropy. It is in principle possible for arbitrarily anisotropic expansion to escape 
390: detection from CMB (considering only the effects from the background), as long as the
391: expansion rates evolve in such a way that $e_z=e_y=0$. In other words, the (background) 
392: quadrupole vanishes, if each scale factor has expanded - no matter how anisotropically -
393: the same amount since the last scattering. An example of such a scenario, derived from the action (\ref{vector}), 
394: is shown in FIG. \ref{tuningpic}.
395: %
396: \begin{figure}[ht]
397: \begin{center}
398: \includegraphics[width=0.4\textwidth,height=0.4\textwidth]{f2.eps}
399: \caption{\label{tuningpic} A minimally coupled vector field satisfying the quadrupole
400: constraint. The solid (black) line is the dark energy density fraction $U$, the dashed (red) line is the 
401: effective equation of state of the universe $w_{eff}$, and the dash-dotted (blue) line describes the 
402: evolution of eccentricity, $E=500e_y^2$. The potential is a double-power law $V = m A^2 + \lambda A^{-4}$. The 
403: dynamics of the field is such that though there are significant anisotropies, the eccentricity at the 
404: present is close to zero. This may be achieved with different power-law potentials, but requires fine-tuning 
405: of the mass scales $m$ and $\lambda$.}
406: \end{center}
407: \end{figure}
408: 
409: \section{SNIa luminosities}
410: 
411: The luminosity-redshift relationship of the SNIa data could be used to probe the possible anisotropies
412: in the expansion history. This is a useful complementary probe since these objects are observed
413: at the $z < 2$ region, whereas CMB comes from much further away at $z \sim 1100$. The luminosity distance 
414: at the redshift $z$ in the direction $\hat{p}$ is now given by \citep{us}
415: \be \label{lumi}
416: d_L(z,{\bf \hat{p}}) = (1+z) 
417: \int_{t_0}^{t(z)}\frac{dt}{\sqrt{\hat{p}_x^2a^2+\hat{p}_y^2b^2+\hat{p}_z^2c^2}}. 
418: \ee
419: To test this prediction with the data, we apply the formula (\ref{z_ellips}) for each
420: observed redshift of a supernova and match its luminosity distance inferred from the observation
421: to the one computed from (\ref{lumi}). In addition, we have to take into account also the angular 
422: coordinates of each individual supernovae in the sky to fix ${\bf \hat{p}}$ for each object.
423: In our analysis we use the GOLD data set \citep{Riess:2006fw}, which consists of five
424: subsets of data\footnote{
425: SNIa
426: Angular coordinates 
427: of each of the 182 GOLD supernovae
428: can be found in \citep{Riess:2006fw,Astier:2005qq} and 
429: http://cfa-www.harvard.edu/ps/lists/Supernovae.html}. We marginalize over the directions
430: in the sky and over the present value of the Hubble constant. 
431: The results are summarized in FIGs. \ref{eoslims} and \ref{eoslims2}. The best-fit anisotropic models 
432: are only slightly preferred over the $\Lambda$CDM, the difference being $\Delta \chi^2 \approx 1$. 
433: Because of the additional parameters in the anistropic models, the reasonable interpretation of these
434: statistics is that the SNIa data favors isotropic expansion.     
435: %
436: \begin{figure}[ht]
437: \begin{center}
438: \includegraphics[width=0.4\textwidth,height=0.4\textwidth]{f3.eps}
439: \caption{\label{eoslims} Limits on the (constant) skewness parameters 
440: of dark energy arising from the SNIa data, when when $\Omega_m=0.3$ and $w=-1$. The contours correspond to
441: 68.3, 90, 95.4 and 99 percent confidence limits. }
442: \end{center}
443: \end{figure}
444: %
445: \begin{figure}[ht]
446: \begin{center}
447: \includegraphics[width=0.4\textwidth,height=0.4\textwidth]{f4.eps}
448: %\includegraphics[width=0.225\textwidth,height=0.5\textwidth]{axis2.eps}
449: \caption{\label{eoslims2} 
450: Constraints arising from the SNIa in the axisymmetric case $\gamma=0$. Inside the darker
451: isosurface, the fit is as good as in the $\Lambda$CDM model, $\chi^2 < 158$. Inside the lighter
452: isosurface, one has $\Delta\chi^2 < 8.02$. One notes that larger skewness $\delta$ would typically be
453: compatible with the SNIa data for phantom equations of state $w<-1$ and large matter densities.}
454: \end{center}
455: \end{figure}
456: 
457: The SNIa data constrains the skewness parameters much looser than the CMB quadrupole, if they are constant.
458: However, in specific models with time-evolving $\delta$ and $\gamma$, the constraints from CMB and from
459: SNIa can be of comparable magnitude and allow anisotropy to an interesting degree.
460: This means that the even if the CMB formed isotropically at early time,
461: it could be distorted by the acceleration of the later universe in such a way 
462: that it appears to us anomalous at the largest scales\footnote{Generating these effects at low redshift has an advantage that  
463: it relaxes constraints which would otherwise come from the CMB polarization \citep{Pontzen:2007ii} and could be strong for a given 
464: temperature anisotropy in isotropizing models because of the significant polarization anisotropy at last scxattering.
465: However, anisotropic dark energy could evade this since the optical depth to $z \sim 1$ is very small.}. The future SNIa 
466: data, with considerably improved error bars
467: %\footnote{
468: (e.g. from the SNAP experiment,  http://snap.lbl.gov/.)
469: %} 
470: on $d_L(z,{\bf \hat{p}})$ might be used to
471: rule out this possibility, and to distinguish whether the possible statistical anisotropy was 
472: already there at last scattering or whether it is due to dark energy.
473: 
474: \section{Conclusions}
475: 
476: In conclusion, dark energy with an anisotropic equation of state 
477: %not only drives the present acceleration, 
478: might be the culprit for both 
479: %understanding the coincidence problem 
480: the cosmic acceleration and the large-angle anomalies in the CMB. This  
481: might also be the key to understand the coincidence problem. The present SNIa data allows anisotropic acceleration, 
482: but SNAP could set things straight about the skewness of dark energy and so of its nature. 
483: Such possibility would  open a completely new window not only on the nature of the CMB anomalies but 
484: also into high energy physics models beyond the usual isotropic candidates 
485: of dark energy such as scalar fields or the cosmological constant. 
486: 
487: \acknowledgments
488: We thank Y. Gaspar, H. Sigbjorn and the anonymous referee for useful comments.
489: TK acknowledges support from the Magnus Ehrnrooth Foundation, the Finnish 
490: Cultural Foundation and EU FP6 Marie Curie Research and Training 
491: Network "UniverseNet" (MRTN-CT-2006-035863). DFM acknowledges support from the A. Humboldt
492: Foundation and the Research Council of Norway through project number 159637/V30. 
493: 
494: \begin{thebibliography}{14}
495: \expandafter\ifx\csname natexlab\endcsname\relax\def\natexlab#1{#1}\fi
496: \expandafter\ifx\csname bibnamefont\endcsname\relax
497:   \def\bibnamefont#1{#1}\fi
498: \expandafter\ifx\csname bibfnamefont\endcsname\relax
499:   \def\bibfnamefont#1{#1}\fi
500: \expandafter\ifx\csname citenamefont\endcsname\relax
501:   \def\citenamefont#1{#1}\fi
502: \expandafter\ifx\csname url\endcsname\relax
503:   \def\url#1{\texttt{#1}}\fi
504: \expandafter\ifx\csname urlprefix\endcsname\relax\def\urlprefix{URL }\fi
505: \providecommand{\bibinfo}[2]{#2}
506: \providecommand{\eprint}[2][]{\url{#2}}
507: 
508: 
509: \bibitem[{\citenamefont{Astier et~al.}(2006)}]{Astier:2005qq}
510: \bibinfo{author}{\bibnamefont{Astier P.}} \bibnamefont{et~al.},
511:   \bibinfo{journal}{Astron. Astrophys.} \textbf{\bibinfo{volume}{447}},
512:   \bibinfo{pages}{31} (\bibinfo{year}{2006}).
513: 
514: \bibitem[{\citenamefont{Armendariz-Picon}(2004)}]{Armendariz-Picon:2004pm}
515: \bibinfo{author}{\bibnamefont{Armendariz-Picon C.}},
516:   \bibinfo{journal}{JCAP} \textbf{\bibinfo{volume}{0407}}, \bibinfo{pages}{007}
517:   (\bibinfo{year}{2004}).
518: 
519: \bibitem[{\citenamefont{Barrow}(1997)}]{Barrow:1997sy}
520: \bibinfo{author}{\bibnamefont{Barrow J.D.}},
521:   \bibinfo{journal}{Phys. Rev.} \textbf{\bibinfo{volume}{D55}},
522:   \bibinfo{pages}{7451} (\bibinfo{year}{1997}).
523: 
524: \bibitem[{\citenamefont{Barrow et~al.}(1985)}]{1985MNRAS.213..917B}
525: \bibinfo{author}{\bibnamefont{Barrow}} \bibnamefont{et~al.},
526:   \bibinfo{journal}{Mon. Not. Roy. Astr. Soc.} \textbf{\bibinfo{volume}{213}},
527:   \bibinfo{pages}{917} (\bibinfo{year}{1985}).
528: 
529: \bibitem[{\citenamefont{Battye and Moss}(2006)}]{Battye:2006mb}
530:   R.~A.~Battye and A.~Moss,
531:   %``Anisotropic perturbations due to dark energy,''
532:   Phys.\ Rev.\  D {\bf 74} (2006) 041301
533:   [arXiv:astro-ph/0602377].
534:   %%CITATION = PHRVA,D74,041301;%%
535: 
536: \bibitem[{\citenamefont{ Behrend, Brown \& Robbers}(2007)}]{back}
537:   Behrend J., Brown I. and Robbers G.,
538:   %``Cosmological Backreaction from Perturbations,''
539:   arXiv:0710.4964 [astro-ph].
540: 
541: \bibitem[{\citenamefont{ Buchert}(2007)}]{backreaction}
542:   Buchert T.,
543:   %``Backreaction issues in relativistic cosmology and the dark energy
544:   %debate,''
545:   AIP Conf.\ Proc.\  {\bf 910}, 361 (2007)
546: 
547: 
548: \bibitem[{\citenamefont{Burd and Lidsey}(1991)}]{Burd:1991ew}
549:   A.~B.~Burd and J.~E.~Lidsey,
550:   %``An Analysis of inflationary models driven by vector fields,''
551:   Nucl.\ Phys.\  B {\bf 351}, 679 (1991).
552:   %%CITATION = NUPHA,B351,679;%%
553: 
554: \bibitem[{\citenamefont{Campanelli et~al.}(2006)}]{Campanelli:2006vb}
555: \bibinfo{author}{\bibnamefont{Campanelli L.}}
556:   \bibnamefont{et~al.}, \bibinfo{journal}{Phys. Rev. Lett.}
557:   \textbf{\bibinfo{volume}{97}}, \bibinfo{pages}{131302}
558:   (\bibinfo{year}{2006}).
559: 
560: \bibitem[{\citenamefont{Copi et~al.}(2006)}]{Copi:2005ff}
561: \bibinfo{author}{\bibnamefont{Copi C.}} \bibnamefont{et~al.},
562:   \bibinfo{journal}{Mon. Not. Roy. Astr. Soc.} \textbf{\bibinfo{volume}{367}},
563:   \bibinfo{pages}{79} (\bibinfo{year}{2006}).
564: 
565: \bibitem[{\citenamefont{Doran and Robbers}(2006)}]{georg}
566:   Doran M. and Robbers G.,
567:   %``Early dark energy cosmologies,''
568:   JCAP {\bf 0606}, 026 (2006)
569: 
570: \bibitem[{\citenamefont{Eriksen et~al.}(2004)}]{Eriksen:2003db}
571: \bibinfo{author}{\bibnamefont{Eriksen H.K.}}
572:   \bibnamefont{et~al.}, \bibinfo{journal}{Astrophys. J.}
573:   \textbf{\bibinfo{volume}{605}}, \bibinfo{pages}{14} (\bibinfo{year}{2004}).
574: 
575: \bibitem[{\citenamefont{Ford}(1989)}]{Ford:1989me}
576:   L.~H.~Ford,
577:   %``INFLATION DRIVEN BY A VECTOR FIELD,''
578:   Phys.\ Rev.\  D {\bf 40} (1989) 967.
579:   %%CITATION = PHRVA,D40,967;%%
580: 
581: \bibitem[{\citenamefont{Hawking}(1968)}]{Hawking:1968zw}
582:   S.~W.~Hawking,
583:   %``On the rotation of the Universe,''
584:   Mon.\ Not.\ Roy.\ Astron.\ Soc.\  {\bf 142} (1969) 129.
585:   %%CITATION = MNRAA,142,129;%%
586: 
587: \bibitem[{\citenamefont{Hinshaw et~al.}(2007)}]{Hinshaw:2006ia}
588: \bibinfo{author}{\bibnamefont{Hinshaw G.}} \bibnamefont{et~al.},
589:   \bibinfo{journal}{Astrophys.J.Suppl.} \textbf{\bibinfo{volume}{170}},
590:   \bibinfo{pages}{288} (\bibinfo{year}{2007}).
591: 
592: \bibitem[{\citenamefont{Jaffe et~al.}(2005)}]{Jaffe:2005pw}
593: \bibinfo{author}{\bibnamefont{Jaffe T.R.}}
594:   \bibnamefont{et~al.}, \bibinfo{journal}{Astrophys. J.}
595:   \textbf{\bibinfo{volume}{629}}, \bibinfo{pages}{L1} (\bibinfo{year}{2005}).
596: 
597: \bibitem[{\citenamefont{Jaffe et~al.}(2006)}]{Jaffe:2005gu}
598: \bibinfo{author}{\bibnamefont{Jaffe T.R.}}
599:   \bibnamefont{et~al.}, \bibinfo{journal}{Astrophys. J.}
600:   \textbf{\bibinfo{volume}{644}}, \bibinfo{pages}{701} (\bibinfo{year}{2006}).
601: 
602: \bibitem[{\citenamefont{Jimenez and Maroto}(2007)}]{BeltranJimenez:2007ai}
603: \bibinfo{author}{\bibnamefont{Jimenez}} \bibnamefont{and}
604:   \bibinfo{author}{\bibnamefont{Maroto}}, \bibinfo{journal}{Phys.Rev.}
605:   \textbf{\bibinfo{volume}{D76}}, \bibinfo{pages}{023003}
606:   (\bibinfo{year}{2007}).
607: 
608: \bibitem[{\citenamefont{Koivisto and Mota}(2006)}]{Koivisto:2005mm}
609:   T.~Koivisto and D.~F.~Mota,
610:   %``Dark energy anisotropic stress and large scale structure formation,''
611:   Phys.\ Rev.\  D {\bf 73} (2006) 083502
612:   [arXiv:astro-ph/0512135].
613:   %%CITATION = PHRVA,D73,083502;%%
614: 
615: \bibitem[\protect\citeauthoryear{Koivisto \& Mota}{2007a}]{tomi1}  Koivisto
616:   T. and Mota D.F.,  (2007a) Phys.\ Lett.\  B {\bf 644}, 104  ; 
617: \bibitem[\protect\citeauthoryear{Koivisto \& Mota}{2007b}]{tomi2} Koivisto
618:   T. and  Mota D.F.,  (2007b) Phys.\ Rev.\  D {\bf 75}, 023518; 
619: 
620: \bibitem[{\citenamefont{Koivisto and Mota}(2008)}]{us}
621: \bibinfo{author}{\bibnamefont{T. Koivisto}} \bibnamefont{and}
622:   \bibinfo{author}{\bibnamefont{D. F. Mota}}
623:   (\bibinfo{year}{2008}), \eprint{Work in progress.}
624: 
625: \bibitem[{\citenamefont{Land and Magueijo}(2005)}]{Land:2005ad}
626: \bibinfo{author}{\bibnamefont{Land}} \bibnamefont{and}
627:   \bibinfo{author}{\bibnamefont{Magueijo}}, \bibinfo{journal}{Phys. Rev. Lett.}
628:   \textbf{\bibinfo{volume}{95}}, \bibinfo{pages}{071301}
629:   (\bibinfo{year}{2005}).
630: 
631: \bibitem[{\citenamefont{Mota et al.}(2007)}]{mota}
632:   Mota D. F., Kristiansen J., Koivisto T. and Groeneboom N.,
633:   %``Constraining Dark Energy Anisotropic Stress,''
634:   arXiv:0708.0830 [astro-ph].
635: 
636: \bibitem[{\citenamefont{Pontzen and Challinor}(2007)}]{Pontzen:2007ii}
637:   A.~Pontzen and A.~Challinor,
638:   %``Bianchi Model CMB Polarization and its Implications for CMB Anomalies,''
639:   arXiv:0706.2075 [astro-ph].
640:   %%CITATION = ARXIV:0706.2075;%%
641: 
642: \bibitem[{\citenamefont{Riess et~al.}(2006)}]{Riess:2006fw}
643: \bibinfo{author}{\bibnamefont{Riess A.G.}} \bibnamefont{et~al.}
644:   (\bibinfo{year}{2006}), \eprint{astro-ph/0611572}.
645: 
646: 
647: \bibitem[{\citenamefont{Schimd et~al.}(2006)}]{uzan}
648: \bibinfo{author}{\bibnamefont{Schimd}} \bibnamefont{et~al.},
649: \bibinfo{journal}{Phys.Rev.D }
650:   \textbf{\bibinfo{volume}{71}}, \bibinfo{pages}{083512}
651:   (\bibinfo{year}{2005}).
652: 
653: 
654: \end{thebibliography}
655: 
656: 
657: \end{document}
658: 
659: