astro-ph0506400/ms.tex
1: \documentclass[12pt,preprint]{aastex}
2: %\documentclass[preprint2]{aastex}
3: \newcommand{\vdag}{(v)^\dagger}
4: \newcommand{\myemail}{skywalker@galaxy.far.far.away}
5: \slugcomment{}
6: 
7: \shorttitle{More evidences for CSL-1} \shortauthors{Sazhin et al.}
8: 
9: \begin{document}
10: 
11: \title{Further spectroscopic observations of the CSL-1 object}
12: 
13: \author{M. Sazhin\altaffilmark{1}, M. Capaccioli\altaffilmark{2,3}, G. 
14: Longo\altaffilmark{3,2,4}}
15: 
16: \and
17: 
18: \author{M. Paolillo\altaffilmark{3,4}, O. Khovanskaya\altaffilmark{1}}
19: 
20:  \affil{1 - Sternberg Astronomical Institute, Moscow State University,
21: University pr. 13, Moscow, RUSSIA}
22:  \affil{2 - INAF - Osservatorio
23: Astronomico di Capodimonte,via Moiariello 16, 80131, Napoli ITALY}
24:  \affil{3 - Department of Physical Sciences, University of Napoli Federico 
25: II,via Cinthia 9, 80126, ITALY}
26:  \affil{4 - INFN, Napoli Unit, ITALY}
27: 
28: \email{sazhin@sai.msu.ru}
29: 
30: 
31: 
32: \begin{abstract}
33: 
34: \noindent CSL-1 is a peculiar object (R.A.$_{2000}= 12^h \ 23^m \ 30"5$; 
35: $\delta_{2000}= -12^{\circ}\ 38' \ 57"0$) which, for its photometric and 
36: spectroscopic properties, is possibly the first case of gravitational lensing 
37: by a cosmic string. 
38: In this paper we present additional evidences, based on medium-high resolution 
39: VLT + FORS1 observations, that the spectra of the two components of CSL-1 are 
40: identical within a confidence level higher than 98$\%$ and the velocity difference of the 
41: two components is consistent with zero. 
42: This result adds further confidence to the interpretation of the system as a 
43: true lens.
44: \end{abstract}
45: 
46: \keywords{gravitational lensing, cosmological parameters, galaxies: peculiar,
47: galaxies: individual(CSL-1)}
48: 
49: \section{Introduction}
50: %
51: As recently stated in the masterly review by \citet{kibble}, the last two 
52: years have seen a renewal of interest in the cosmological role of cosmic 
53: strings, after more than a decade of relative quiescence: 
54: an interest which was mainly triggered by the discovery of
55: the unusual object CSL-1 (R.A.$_{2000}= 12^h \ 23^m \ 30"5$;
56: $\delta_{2000}= -12^{\circ}\ 38' \ 57"0$) \cite[][hereafter Paper I]{csl1} 
57: in the OAC-DF
58: (Osservatorio Astronomico di Capodimonte - Deep Field; \citealp{cap2})
59: and by the indirect evidence
60: obtained by \cite{schild} from the luminosity fluctuations of
61: the quasar $Q0957+561 \ A,B$.
62: 
63: CSL-1 (see Fig.\ref{CSL1}) is a double source laying in a
64: low density field.
65: In the original images the components, 1.9 arcsec apart, appeared to be 
66: extended and with roundish and identical shapes. 
67: By low resolution spectroscopy we learned that both components are at a redshift 
68: of $0.46\pm 0.008$, and by photometry (both global properties and luminosity 
69: profiles) that they match the properties of two giant elliptical galaxies. 
70: Detailed analysis showed that the spectra of the two components were
71: identical at a 99.9\% level. 
72: Such a conclusion, however, was hindered both by the limited wavelength range 
73: spanned by the spectra, and by their relatively low signal-to-noise ratio.
74: 
75: As discussed in Paper I, the only possible explanation of the CSL-1
76: properties is: {\it i)} either an unlikely chance alignment of two
77: giant ellipticals at the same redshift and with very similar spectra, or
78: {\it ii)} a gravitational lensing phenomenon. 
79: But in this second case, due to the lack of asymmetry in the two images, the 
80: lens could not be modelled with the standard lensing by a massive compact 
81: source. 
82: Actually, the usual gravitational lenses, i.e. those formed by bound clumps of 
83: matter, always produce inhomogeneous gravitational fields which distort the 
84: images of extended background sources \cite[cf.][]{gl,kee}. 
85: The detailed modelling of CSL-1 proved that the two images were virtually 
86: undistorted (see Paper I for details).
87: The only other explanation left in the framework of the gravitational lensing
88: theory was that of a lensing by a cosmic string.
89: 
90: In Paper I we indicated two possible {\it experimenta crucis}: {\it i)} the 
91: detection of the sharp edges at faint light levels, since this is the signature 
92: expected from the lensing by a cosmic string, and {\it ii)} the detection of a 
93: small amplitude but sharp discontinuity in the local CMB \citep{gangui}. 
94: The first test calls for high angular resolution and deep observations with HST; 
95: time has already been allocated but the observations have not yet been 
96: performed. 
97: The second test was attempted using the preliminary release of the WMAP data 
98: \citep{kai84, WMAP}. 
99: By careful processing, a $1 \sigma$ positive detection at the position of CSL-1 
100: was found. 
101: Even though such a detection seems to imply the unrealistic speed of $0.94 \ c$ 
102: for the string, the authors pointed out that both the low angular resolution and 
103: the low S/N ratio could prevent the detection of the expected signature.
104: 
105: Another test was proposed by several authors \citep{string1, string2, string3, string4}, 
106: and more recently in \citet{string5}.
107: It is based on the fact that the alignment of the background object (a galaxy) 
108: inside the deficit angle of the string is a stochastic process determined by the 
109: area of the lensing strip and by the surface density distribution of the
110: extragalactic objects which are laying behind the string. 
111: All the lensed objects will fall inside a narrow strip defined by the
112: deficit angle computed along the string pattern. 
113: A preliminary investigation of the CSL-1 field showed a significant excess of
114: gravitational lens candidates selected on the bases of photometric
115: criteria only \citep{saz04}. 
116: Spectroscopic observations are being obtained for a first set of candidates and 
117: will be discussed elsewhere.
118: 
119: \noindent In this paper we address on firmer grounds the issue of the 
120: gravitational lens nature of CSL-1 using intermediate resolution spectra obtained 
121: at the ESO Very Large Telescope + FORS 1 spectrograph.
122: 
123: \section{The data and data reduction}
124: %
125: New spectra of CSL-1 were obtained in March 2005 at VLT\footnote{The Very Large 
126: Telescope is operated by the European Southern Observatory and is located at 
127: Mount Paranal in Chile; http://www.eso.org/paranal/} using the FORS 1
128: spectrograph under Director's Discretionary Time (proposal 274.A-5039).
129: 
130: The spectra were acquired on March 15-19 2005
131: in the FORS 1 long slit configuration 600V+GG435 ($\lambda/\Delta\lambda=990$ at central 
132: wavelength), setting the slit of the spectrograph across the centers of the two components of 
133: CSL-1. The observations were split in several exposures, to prevent saturation and to 
134: allow for a better removal of the bad pixels and cosmic rays.
135: The risk of cross contamination between the spectra of the two components was 
136: minimized by retaining only the 6 frames with an average Signal to Noise (S/N) ratio of $\sim 
137: 12$ and a PSF FWHM$<1.0"$\footnote{The Point Spread Function represents the image of a pointlike object. 
138: Its Full Width at Half Maximum measures the level of blurring due to atmospheric and instrumental factors.}, for a total exposure time of 4740 sec. 
139: 
140: During observations a short $R$-band exposure was also obtained to check the 
141: pointing of the instrument. 
142: Owing to the excellent seeing conditions, this image could be used to test the
143: extended nature and the similarity of the light profiles of the two images. 
144: The dot-dashed line (with no data symbols) in Figure~\ref{lightprofile} shows 
145: the light profile of an unresolved star present in the field, compared to the 
146: light profiles of the two images of CSL-1, which appear to be clearly resolved 
147: and identical within the errors. A de Vaucouler fit yields $r_e\simeq 1.6"$.
148: 
149: The spectral data were reduced through standard MIDAS procedures
150: and, after bias subtraction, flat fielding, wavelength calibration,
151: and sky subtraction, the spectra were re-aligned to correct for
152: dithering and stacked. Since the shift values were set to an integer
153: number of pixels, no re--sampling was applied, in order not to
154: affect the noise statistics. The stacking was performed using a
155: simple median filter to reject cosmic rays.
156: 
157: The spectra of each component was extracted using a 5 pixel strip (1
158: pix=0.21") centered on the emission peak with the purpose of maximizing the
159: S/N ratio. In a similar way, the background counts were extracted in
160: two stripes located 40 pixels from each component so to
161: measure the local background while minimizing the contribution from the
162: source. The error on the spectral counts was calculated with the
163: following expression:
164: %
165: $$\sigma(ADU)=\sqrt{\sigma_{bkg}^2 + \frac{N(ADU)}{n_{exp}
166: \times g}}$$
167: %
168: where: $N(ADU)$ are the source counts measured along the
169: spectrum, $n_{exp}$ are the number of median averaged exposures, $g$
170: is the instrumental gain, and $\sigma_{bkg}$ represents the background
171: r.m.s. measured over $5\times 20$ pixels centered at each wavelength.
172: By folding the images we estimate an average cross-contamination of $7\%\pm 1\%$.
173: The resulting spectra and their ratio are shown in
174: Figure~\ref{spectra}. No flux calibration was applied to the data.
175: It has to be stressed that the narrow spikes visible in the Figure
176: are residuals left after the removal of bright sky lines.
177: 
178: The spectra of the two components turned out to be identical at visual 
179: inspection. In fact Pearson's, Spearman, and Kendall correlation tests indicate
180: correlation coefficients of $0.96$, $0.94$ and $0.94$ respectively, with a significance
181: of $>99.9\%$ in all cases.
182: The degree of similarity was further quantified by running a $\chi^2$ test
183: both on the whole wavelength range\footnote{The regions affected by sky lines subtraction residuals 
184: were excluded from the test.} and on the most prominent spectral absorption features, namely 
185: Ca II ($H$, $K$), H$\beta$, $H\delta$, $H\gamma$ lines and $G$ band. The
186: test yields $\chi^2_\nu=1.03$ implying that the two spectra are 
187: consistent within $<2\sigma$ (80\%); we also compared the distribution of the 
188: observed differences to the frequencies expected in the case of pure gaussian noise, finding that the two 
189: are consistent at the 95\% level and that there is no deterministic part in the residuals. 
190: An even better agreement ($1\sigma$) is found for the individual absorption features.
191: To further check if the observed consistency can be due to the known similarity of
192: early-type galaxies we repeated the test on a sample of spectra extracted from the 
193: SDSS\footnote{Sloan Digital Sky Survey, data release 4: http://www.sdss.org} {\it Luminous Red Galaxies}, chosen to have a redshift difference and S/N comparable with the CSL1 data. We performed 2000 comparisons obtaining that $<2\%$ of the examined SDSS spectra are as consistent as the spectra of the two CSL1 components. 
194: 
195: A cross correlation test based on the spectral 
196: lines mentioned above yelds a velocity difference between the two components of 
197: $\Delta v=14\pm 30$ Km s$^{-1}$. If however we exclude the $H\beta$ line which is 
198: affected by residual instrumental effects this figure reduces to $0 \pm 20$ Km s$^{-1}$. 
199: 
200: 
201: \section{Conclusion}
202: 
203: The similarity of the spectra of the two components of CSL-1 and the
204: zero velocity shift between them strongly support the
205: interpretation of CSL-1 in terms of gravitational lensing.
206: The data obtained so far do not allow to completely rule
207: out the possibility of a chance alignment of two giant ellipticals
208: but the new data presented in this paper make such an
209: alignment very unlikely. 
210: In the case of chance alignment, in fact, the two images of CSL-1
211: would correspond to two giant ellipticals with identical shapes and
212: spectra, placed at the same redshift. 
213: The probability of finding two ellipticals of $M_R=-22.3$ within 2" (20 kpc) and with a radial distance 
214: $<1$ Mpc ($2\sigma$ upper limit) is $P\sim 1.5\times 10^{-15}$,
215: accounting for clustering effects \citep[e.g.][]{zeh}\footnote{$P=\int_{V_1,V_2} N^2_{gal}[1+\xi(r)]dV_1 dV_2$ were 
216: $N_{gal}$ is the space density of elliptical galaxies, $\xi(r)$ is the galaxy correlation function, 
217: $V_1$ is the volume enclosing the two galaxies and $V_2$ is the volume of the survey.}. 
218: Integrating over the volume sampled by the OAC-Deep Field for a galaxy of the same magnitude as CSL-1 we calculate that we expect to find $\sim 9\times 10^{-4}$ pairs in the whole survey. Including the spectral similarity we obtain an upper limit of $P<2\times 10^{-5}$.
219: As it was already mentioned in
220: Paper I, this could still be explained if CSL-1 belonged to a rich cluster with two central
221: dominant galaxies, but this is not the case. Careful inspection of the
222: CSL-1 field shows in fact that it is a rather isolated object with
223: no other nearby galaxies of comparable brightness; furthermore the 
224: velocity difference measured from the two spectra is much smaller 
225: that the one expected in a rich cluster ($\Delta v\sim 300$ Km s$^{-1}$).
226: 
227: In the gravitational lensing scenario, as already stated in Paper I, the observed phenomenology
228: cannot be understood in terms of lensing by compact clumps of matter such as,
229: for instance, a Singular Isothermal Sphere model or any other model
230: listed in the C.R. Keeton's Lens Modeling Software \citep{kee}. 
231: The only possible type of lens which can produce a morphology similar to
232: that observed in CSL-1 seems to be a cosmic string \citep{csl1}. 
233: 
234: Lensing by a cosmic string seems capable to
235: explain all the observational evidences gathered so far and deserves
236: further investigation. Cosmic strings were predicted by 
237: \cite{kib76} and their role in cosmology has been
238: extensively discussed by \cite{zel} and 
239: \cite{vil}. Recent work \citep{kibble, pol, dav05} has also shown
240: their relevance for both fundamental physics and cosmology. In
241: particular it has become apparent that the detection of a cosmic
242: string would lead to a direct  measure of the energy scale of
243: symmetry breaking in GUT theories. If we assume that CSL-1 is
244: produced by a cosmic string, its measured properties would imply a
245: linear density of the string of order of $G\mu \approx 4\cdot
246: 10^{-7} $ and a corresponding energy scale of GUT of $\sim 10^{15}$
247: GeV \citep{kib76, pdg}.
248: 
249: Hopefully the question on the nature of CSL1 will soon be answered by
250: our HST observations approved in Cycle 14 to carry out the 
251: test proposed by the authors in Paper I and which will allow to verify
252: the cosmic string hypothesis on firmer grounds.
253: 
254: \section*{Acknowledgments}
255: 
256: M.V.Sazhin acknowledges the INAF-Capodimonte Astronomical Observatory for
257: hospitality and financial support. O. Khovanskaya acknowledges the
258: Department of Physics of the University Federico II in Naples for
259: financial support. The authors wish to thank C. Cezarsky, Director General
260: of the European Southern Observatory for allocating Director's
261: Discretionary Time to this project. The work was also supported by
262: the Russian Fund of Fundamental Investigations No. 04-02-17288. We thank the anonymous 
263: referees for the helpful suggestions.
264: 
265: 
266: \begin{thebibliography}{}
267: 
268: \bibitem[Alcal\'a et al.(2004)]{cap2}
269: Alcal\'a J.M., Pannella M., Puddu E., et al. 2004, Astron. and
270: Astrophys., 428, 339
271: 
272: \bibitem[Bernardeu \& Uzan(2001)]{string3}
273: Bernardeau F., \& Uzan J.-P. 2001, Phys.Rev D, 63, 023004, 023005
274: 
275: \bibitem[Davis \& Kibble(2005)]{dav05}
276: Davis A.-C., \& Kibble T.W.B., 2005, http://ArXiv.org/hep-th/0505050
277: 
278: \bibitem[de Laix et al.(1996)]{string4}
279: de Laix, A.A., Vachaspati, T. 1996, Phys.Rev. D 54, 4780, 1996.
280: 
281: \bibitem[Gangui et al.(2001)]{gangui}
282: Gangui A., Pogosian L., Winitzki S. 2002, NewAR, vol.46, 11, 681 
283: 
284: \bibitem[Hindmarsh(1990)]{string2}
285: Hindmarsh. A.,in: The Formation and Evolution of Cosmic Strings,
286: ed. by G.Gibbons, S.W.Hawking, and T.Vachaspathi.  Cambridge
287: Univ.Press., Cambridge, 1990, p.527.
288: 
289: \bibitem[Huterer \& Vachaspati(2003)]{string5}
290: Huterer, D., Vachaspati, T. 2003, PhRevD, 68, 4
291: 
292: \bibitem[Lo \& Wright(2005)]{WMAP}
293: Lo A.S., Wright E.L., 2005, http://ArXiv.org/astro-ph/0503120
294: 
295: \bibitem[Kaiser \& Stebbins(1984)]{kai84}
296: Kaiser N. and Stebbins A., 1984, Nature, v.276, p.591.
297: 
298: \bibitem[Kochanek et al.(2004)]{kee}
299: Kochanek C.S., Falco E.E., Impey C., LeharJ., McLeod B., Rix H.-W., 2004;
300: http://cfa-www.harvard.edu/castles/
301: 
302: \bibitem[Kibble(1976)]{kib76}
303: Kibble T.W.B., 1976, J.Phys.A: Math \& Gen. v.9, 1387.
304: 
305: \bibitem[Kibble(2004)]{kibble}
306: Kibble T.W.B., 2004, http://ArXiv.org/astro-ph/0410073, in Proc. of COSLAB 2004
307: 
308: \bibitem[Particle data group(2004)]{pdg}
309: Particle Physics booklet. Elsevier, 2004.
310: 
311: \bibitem[Polchinsky(2004)]{pol}
312: Polchinski J., 2005, AIP Conf.Proc. 743, 331; http://ArXiv.org/hep-th/0410082
313: 
314: \bibitem[Sazhin et al.(2003)]{csl1} Sazhin, M., Longo, G.,
315: Capaccioli, et al. 2003, MNRAS, 343, 353
316: 
317: \bibitem[Sazhin et al.(2005)]{saz04} Sazhin, M., Khovanskaya O.,
318: Capaccioli, et al. Astronomy Letters, 31, 73; http://ArXiv.org/astro-ph/0406516
319: 
320: \bibitem[Schild(2004)]{schild}
321: Schild R., Masnyak I.S., Hnatyk B.
322: I., Zhdanov V.I., 2004, AA, 422, 474
323: 
324: \bibitem[Schneider et al.(1992)]{gl}
325: Schneider P., Ehlers J., Falco E.E., 1992, Gravitational Lenses,
326: Springer, Heidelberg.
327: 
328: \bibitem[Vilenkin(1981)]{vil}
329: Vilenkin A., 1981, Phys.Rev.,D, 23, 852
330: 
331: \bibitem[Vilenkin \& Shellard(1994)]{string1}
332: Vilenkin A., Shellard E.P.S, 1994, Cosmic strings and other
333: topological defects. Cambridge Univ.Press., Cambridge.
334: 
335: \bibitem[Zehavi et al.(2005)]{zeh}
336: Zehavi, I., et al. 2005, ApJ, 630, 1
337: 
338: \bibitem[Zeldovich(1980)]{zel}
339: Zeldovich, Ya.B., 1980, MNRAS, 192, 663
340: 
341: \end{thebibliography}
342: 
343: \clearpage
344: 
345: \begin{figure}
346: \includegraphics*[width=7cm]{f1.eps}
347:  \caption{The CSL-1 field with the slit position drawn upon it.}\label{CSL1}
348: \end{figure}
349: 
350: 
351: \clearpage
352: 
353: \begin{figure}
354: \includegraphics*[width=8cm]{f2.eps}
355:  \caption{The integrated light profiles for the two components of CSL-1
356: (solid and dashed line) compared to the light profile of an unresolved source
357: (dot-dashed line). Galaxy profiles were extracted avoiding the region of overlap,
358:  i.e a sector of $60^\circ$ toward the other component. Errorbars represent $1\sigma$
359:  uncertainties.}
360: \label{lightprofile}
361: \end{figure}
362: 
363: 
364: \clearpage
365: 
366: \begin{figure*}[t]
367: \includegraphics[width=12cm,angle=90]{f3.eps}
368:  \caption{Upper panel: the spectra of the two components of CSL-1.
369:  Lower panel The ratio between the two spectra in the upper panel.
370:  The $1$ and $3\sigma$ limits are shown
371:  as dark and light grey shaded regions.}\label{spectra}
372: \end{figure*}
373: 
374: \end{document}
375: 
376: