1: \documentstyle[psfig,11pt]{article}
2:
3: \topmargin=-1cm
4: \oddsidemargin=5mm
5: \textwidth=16.5cm
6: \textheight=23cm
7: \sloppy
8:
9: \begin{document}
10:
11: \centerline{\large\bf A DIFFERENTIAL METHOD OF SEARCH FOR}
12: \vskip 2mm
13: \centerline{\large\bf THE CMBR SPECTRAL-SPATIAL FLUCTUATIONS}
14:
15: \bigskip
16:
17: \centerline{\bf Victor K. Dubrovich$^1$, Anisa T. Bajkova$^2$}
18:
19: \bigskip
20:
21: \centerline{$^1$\it Special Astrophysical Observatory RAS, e-mail: dubr@MD1381.spb.edu}
22:
23: \medskip
24:
25: \centerline{$^2$\it Institute of Applied Astronomy RAS, e-mail: bajkova@quasar.ipa.nw.ru}
26:
27: \bigskip
28: \begin{abstract}
29: The CMBR spectral-spatial fluctuations (SSF) formed in early Universe
30: during the Dark Ages are considered. Main attention is focused on the
31: narrow-band spectral properties of the SSF. Based on these properties we
32: propose to use a differential method in order to search for these fluctuations.
33: Description of the method is given.
34: \end{abstract}
35:
36: \section{Introduction: Structure Formation and Big Bang Cosmology}
37:
38: The Big Bang theory for the expanding Universe is now well
39: established. One of the main observation effect caused by this expansion
40: is the redshift of photons, expressed as $\lambda_{obs}=(z +1)\cdot
41: \lambda_{em}$,
42: where $\lambda_{em}$ is the emitted or rest wavelength of the photon, and $z$
43: is the redshift.
44:
45: One early moment of particularly great
46: importance occurs at a redshift $z\approx 1100$, corresponding to a time
47: $3\cdot 10^5$ yr.
48: after the Big Bang. This is the moment observed when we study the
49: distribution of the cosmic microwave background radiation (CMBR), heat left
50: over from the Big Bang. The importance of this surface is explored in reviews
51: such as Silk, 1994. In effect, the distribution of intensity of CMBR
52: on the sky provides information about the large-scale distribution
53: of matter at a very early time in the history of the Universe. For this
54: reason, both NASA and the European Space Agency have attached high priority
55: to the mapping of this surface using the CMBR. We know from earlier work
56: that it is largely featureless, that is that the distribution of matter was
57: quite homogeneous early in the history of the Universe. However, the COBE
58: satellite (Bennett et al., 1993), and recently BOOMERANG (Lange et al.,
59: 2001) and MAXIMA (Hanany et al., 2000) revealed
60: small fluctuations in the CMBR temperature, corresponding to perturbations
61: in the density of matter of order $10^{-3}$. These are the seeds of all the
62: structure we see in the Universe today. To summarize, the Big Bang itself
63: was quite homogeneous, but very low amplitude fluctuations in density were
64: present at an epoch of $3\cdot 10^5$ yr.
65:
66: On the other hand, the present Universe is exceedingly lumpy~ ---~ there
67: are large-scale structures ranging in size from galaxies to clusters of
68: galaxies and even larger systems. Recent optical and radio studies, for
69: instance by the Hubble Space Telescope, have established that structure is
70: present in the Universe back to redshifts at least $z=5$.
71:
72: It follows from these two sets of observations that much of the
73: structure currently seen in the Universe first formed in the redshift
74: interval 6--1000. It is a deeply interesting question to know when and how
75: the large-scale systems we see in the Universe formed. Unfortunately, there
76: are essentially no observations available in that redshift interval,
77: corresponding to the time period $3\cdot 10^5$ to $10^9$ years after the Big
78: Bang (although there are some hints from Lanzetta et al., 1999 of the existence
79: of galaxies at $z>6$). Perhaps for this reason, this phase in the history
80: of the Universe is known as the "Dark Ages" of the Universe.
81:
82: \section{The Dark Ages}
83:
84: While we have no direct observations of objects or proto-objects in
85: the Dark Ages, we may make some inferences. First, sometime in that
86: interval, the small amplitude, linear perturbations in the density of
87: matter must have become nonlinear, since the densities within galaxies
88: observed even at $z=5$ are much higher than the background density. The
89: physics of nonlinear gravitational collapse are much more complicated and
90: rich than the simpler physics of linear gravitational contraction. In other
91: words, a good deal of interesting astrophysics certainly occurred during
92: the Dark Ages. Also, some very approximate constraints can be placed on
93: events during the Dark Ages by looking at integrated background light. For
94: instance, there are upper limits on the energy release at that range of
95: redshifts established by observations by the COBE satellite (see Hauser et
96: al., 1998; Dwek et al., 1998 and Haarsma and Partridge, 1998).
97:
98: Given the importance of questions about the formation of structure
99: in our Universe, it is obviously important to explore any means to increase
100: our knowledge about the Dark Ages of the Universe. The interaction of the
101: CMBR with molecular species created in the early Universe offers such a
102: possibility. In particular, the interaction of the CMBR with lumpy matter
103: containing simple molecules will introduce wavelength dependent
104: anisotropies in the CMBR: these are the spectral-spatial fluctuations
105: (SSF) described in detail by Dubrovich (Dubrovich, 1977, 1994, 1997). The
106: amplitude of the SSF signals is not
107: large, but current and planned observation programs offer the possibility
108: that these clues to the formation of structure early in the Universe can be
109: detected.
110:
111: \section{Theory of Spectral-Spatial Fluctuations}
112:
113: During the Dark Ages, there is a period when the material contents
114: of the Universe are present in the form of atoms and molecules, with low
115: amplitude density fluctuations and a low number of free electrons. The main
116: "objects" at this epoch are diffuse and low-contrast protogalaxies. The
117: temperatures of matter and radiation (the CMBR) are practically equal until
118: the redshift drops to 150.
119:
120: It is possible to generate energy within these proto-objects as a result of
121: heating due to their collapse and consequent radiative cooling. But the low
122: matter density and low ionization lead to very low luminosity in free-free
123: emission or atomic transitions. Another mechanism which can produce
124: observation effects is the Doppler effect by peculiar motion of
125: a protogalaxy with a velocity $V_p$ (Dubrovich, 1977, 1997; Maoli et
126: al., 1996).
127: Note that in this case there is no need for an
128: internal energy source in the proto-object. The value of the distortion
129: produced in the CMBR is:
130: \begin{equation}
131: \delta T/T = (V_p/c) \tau,
132: \end{equation}
133: where $c$ is the speed of light and $\tau$ is the
134: optical depth of the proto-object ($\tau<1$). In the case of resonant transitions (line
135: emission) in molecules, the optical depth $\tau$ can be strongly frequency
136: dependent, having a high value in a narrow bandwidth around each resonant
137: frequency.
138: But in any cases it is possible to make one general affirmation:
139: due to this mechanism one can obtain only the value
140: $\delta T/T<V_p/c<10^{-3}$.
141:
142: Several authors have investigated the most probable primordial
143: molecules and the effects they create. The expected abundance of these
144: simple molecules in the early Universe was discussed by Puy et al., 1993;
145: Lepp and Shull, 1984; Palla et al., 1995; Maoli et al., 1996 and Stancil et
146: al., 1996.
147:
148: We now show that these primordial molecules, which may produce
149: observable effects in the CMBR, do trace the initial large-scale
150: distribution of matter (1). The rate of molecule formation depends on many
151: physical parameters such as the electron density, $n_e$. Free electrons
152: catalyze the formation of neutral molecules, for instance $H_2$; in turn, the
153: abundance of other molecules are strongly correlated with $H_2$ concentration.
154: It follows that fluctuations in the matter density and $n_e$ lead to nonlinear
155: enhancement of the optical depth in molecular lines.
156: The kinetic temperature of matter also is involved after $z\approx 150$. It is
157: higher in the denser regions, leading to a shift of equilibrium in the ratio
158: of $H^{+}/H_2$, $HeH^{+}/H_2$ and $LiH/H_2$
159: towards higher values (2). However, the peculiar velocity continues to be
160: the dominant effect, and this results from hydrodynamic motions of matter
161: strongly correlated with density fluctuations (e.g., Vishniac, 1987).
162: Later, nonlinear stages of the evolution of proto-objects are characterized
163: by the high temperature of matter and high ionization due to compression
164: and shocks. This lies in the interval $50>z>5$. Ionizing photon flux may
165: be sufficient to ionize hydrogen, but not strong enough to render the
166: protogalaxies visible in the optical or the radio range. The Doppler
167: mechanism still dominates. The $He H^{+}$ molecule will play a particularly
168: important role. The rest wavelength of its first rotational transition is
169: 0.149 mm, with other lines at 0.075 mm and 0.05 mm. At $z=5$, these become
170: 0.9 mm, 0.45 mm and 0.3 mm, respectively.
171:
172: We note that $He H^{+}$ will be present wherever high temperature,
173: compressed matter is present. In addition, the fact that the effect on the
174: CMBR is a resonant one with $t$ depending strongly on frequency means that
175: observations at any fixed wavelength are strongly dependent on the
176: redshift. Thus we can obtain a complete 3-dimensional picture of structural
177: formation including the radial dimension, practically impossible to obtain
178: in any other way.
179:
180: The final stage of the evolution of matter is the formation of the
181: first stars. The starburst phenomena will lead to the synthesis of heavy
182: elements, especially $C, N$ and $O$. This allows for the formation of molecules
183: such as $OH^{+}, OH, CH$ and $CH^{+}$, etc. At the same time, some matter will
184: be accelerated to high speeds due to the expulsion of shells in the starburst
185: phenomena. We will investigate whether these molecules also produce
186: observation SSF effects.
187:
188: \section{Spectral properties of SSF}
189:
190: Except of pure scattering the luminescent transformation of some kind of
191: superequilibrium photons is possible as well as the combination of both
192: mechanisms (Dubrovich, 1977, 1994, 1997, 1997a, 1999, Dubrovich et al., 1995).
193: In all these cases the essential peculiarity of the fluctuations, namely
194: strong frequency dependence of the effect, becomes apparent.
195: So, if an object contains gas of any molecules and has a redshift
196: $z$, it is seen only at frequencies $\nu_i$:
197: \begin{equation}
198: \nu_i = \nu_{oi}/(1+z),
199: \end{equation}
200: where $\nu_{oi}$ is a discrete set of the molecule's rest transition frequencies.
201: It means that an object can be observed at different but discrete
202: frequencies. And vice versa, if some object is seen at the given
203: frequency $\nu$, it means that this object can have any of the redshifts $z_i$:
204: \begin{equation}
205: z_i = \nu_{oi}/\nu - 1.
206: \end{equation}
207:
208: This uncertainty of $z_i$ can be cleared up, for instance, by such a way as for
209: distant quasars: it is necessary to obtain spectrum of interested object
210: wide enough. If we suppose that this line caused by any rotational transition
211: in simple two-atomic molecules it is necessary to get the spectrum interval
212: from $\nu_1$ to $\nu_2=2\cdot \nu_1$. In this case we can guarantee, that all
213: possible
214: variants of rotational number $J=0,1,2,...$ corresponded to this transition could be
215: known.
216: Indeed, the frequency of rotational transition (from level with rotational
217: number $J$ to level $J+1$, for scattering) $\nu_J$ is proportional to
218: $J+1$. So the
219: frequencies ratio of two nearest observed lines will be $(J+2)/(J+1)\le 2$. The
220: maximum of this ratio corresponds to the lines from $J=0$ to $J'=1$ and
221: from $J'=1$
222: to $J''=2$. If we see the ratio less then 2 we can calculate $J$ which
223: is the number of the low rotational level. For example, if this ratio is
224: 1.5 it means that we see transitions from $J=1$ to $J'=2$ and from
225: $J'=2$ to $J''=3$.
226:
227: One of the consequences of (2), (3) is that the point object with redshift
228: $z_1$ visible at frequency $\nu_1$ is not visible at another frequency
229: $\nu_2$, if the following relation is realized:
230: \begin{equation}
231: \nu_1<\nu_2;~~\nu_1 =\nu_{oi}/(1+z_1);~~\nu_2<\nu_{o(i+1)}/(1+z_1).
232: \end{equation}
233:
234: Here we suppose that the size $L$ of the object is proper small.
235: More exact condition for $\Delta\nu$, taking into account the finite
236: size of an object, will be done below.
237: At the same time at the frequency $\nu_2$ we will see the objects with
238: redshift $z_2$ in accordance with (3), which may be different from the
239: seen at the frequency $\nu_1$.
240:
241: The angular size $\theta$ and the frequency bandwidth $\Delta\nu$, in which
242: the given object can be observed, depends on its linear size $L$.
243: For small values of $z$ the angular size of an object (when the linear
244: size is constant and equal to $L$) is decreasing with moving away,
245: but for $z$ larger than some value, it is increasing.
246: The modern observation data select the model of plane expanding Universe
247: with cosmological constant $\Lambda$.
248: Exact relations between $\theta$ and $z$ in this model contain the
249: parameters $\Omega_m$ and $\Omega_{\Lambda}$, which are the ratios of the
250: matter $n_m$ and "vacuum" densities to the critical density $n_c$ respectively
251: ($\Omega_m+\Omega_{\Lambda}=1$) and look like (Sahni et al., 1999):
252: $$
253: \theta=(H_0L/c)(1+z)/\phi(z),
254: $$
255: $$
256: \phi(z)=\int_0^z[\Omega_m(1+z)^3+\Omega_{\Lambda}]^{-1/2}dz.
257: $$
258:
259: For large $z$ and $H_0=60$km/s$\cdot$Œps$\cdot$$h_{60}$,
260: $\Omega_m=0.3$ and $L$ in Mps we obtain
261: $$
262: \theta \approx 15^{\prime\prime}h_{60}L(1+z).
263: $$
264:
265: On the other hand, front and back edges of molecular cloud are at
266: different distances from us, i.e. have different $z$.
267: The value of this difference of $z$ is connected with $L$
268: in the following way (Sahni et al., 1999):
269: $$
270: \Delta z/(1+z) = (H_0 L/c) [\Omega_m(1+z)^3+\Omega_{\Lambda}]^{1/2}.
271: $$
272:
273: The fact that the cloud occupies the interval of redshifts $z$
274: means, that if it radiates or reflects the radiation locally in sufficiently
275: narrow lines, then all the radiation occupies the frequency interval
276: $\Delta\nu$, moreover
277: $$
278: \Delta\nu/\nu=\Delta z/(1+z).
279: $$
280: Consequently,
281: $$
282: (\Delta\nu/\nu)\approx \theta\phi(z) \cdot [\Omega_m(1+z)^3+\Omega_{\Lambda}]^{1/2}(1+z)^{-1}.
283: $$
284: Numerically for large $z$ and $\Omega_m=0.3$ we obtain
285: \begin{equation}
286: (\Delta\nu/\nu)\approx 1.310^{-3}(\theta/1^{\prime}) \cdot [\phi(z)/\phi(20)](1+z)^{1/2}.
287: \end{equation}
288:
289: If we simultaneously observe with two receivers with frequency
290: difference $\Delta\nu$, the fluctuations with the size larger than $\theta$
291: are seen at the both. If the size of the fluctuations is smaller, they
292: are seen only with one of the receivers, another receiver will detect
293: only a background noise (Dubrovich, 1982).
294: Thus the correlation function of these two observations will be
295: equal to within a noise to zero at small scales and different of zero
296: at scales greater than certain.
297: The presence of clear break is possible only if all the objects are
298: of equal size $L$.
299: Practically there is some spectrum of scales.
300: The effect considered shows that the initial spectrum can be
301: significantly distorted.
302: More full investigation of this problem stands out of this paper.
303: But from the said already it is seen that the analysis of SSF will
304: give additional information on the matter parameters and redshifts $z$
305: of basic proto-objects.
306:
307: Scattering by molecules in several lines leads to superposition of
308: images of different objects of different $z$.
309: As shown in Dubrovich (1977), the cosmological molecules $LiH$ and $HD^{+}$
310: can exist at $50<z<200$.
311: The interval of wavelengths for $LiH$ spectral lines in this condition
312: lies between 13 cm for $z=200$ and $i=1$ and $0.5$ cm for $z=50$ and $i=6$.
313: For $HD^{+}$ it spreads from 4.4 cm to 0.2 cm.
314: This leads to increasing of fluctuations number at large frequencies.
315: The power spectra obtained in cm and mm ranges must differ by the increase
316: of the part of small scale fluctuations in the second case.
317:
318: The estimations and equations mentioned above are correct for objects
319: freely expanding over the Universe in accordance with the Hubble law.
320: But really at the sufficiently late stages the deceleration of their
321: expansion takes place and contraction begins due to self-gravitation.
322: In this case the line's width can be considerably smaller, and the
323: amplitude larger (Zel'dovich, 1987).
324:
325: \section{The differential method for detecting the SSF}
326:
327: The differential method can be considered as an alternative method to the
328: correlation function analysis indicated above.
329: In this case the subject of interest is actually the first derivative
330: of spatial distortions with respect to the frequency.
331:
332: The method used based on the analysis of a difference of two CMBR
333: temperature maps
334: observed at different frequencies and reduced to one beam shape.
335: Obviously, such a difference map contains information only on the
336: secondary fluctuations because the primary CMBR fluctuations present in
337: both maps will be eliminated in subtraction result due to their black-body
338: spectrum nature.
339:
340: Let the difference of the CMBR observation frequencies be equal to
341: $\Delta\nu_1$.
342: Let the limiting angular fluctuation size $\theta_1$ correspond to this
343: frequency difference in accordance with equation (5).
344: Obviously, the fluctuations of size larger than $\theta_1$ will be seen
345: at both maps. After subtraction of one map from the other the
346: fluctuations of size larger than $\theta_1$ will be mutually compensated
347: and the remainder map will contain only the fluctuations smaller
348: than $\theta_1$.
349: Evidently, this fact can be seen also from angular power spectrum of
350: this map which is close to zero at frequencies with multipoles $l<l_1$
351: and different from zero at $l>l_1$, where $l_1\approx 1/\theta_1$ is the
352: break of the spectrum.
353: Now let the difference of two frequencies be $\Delta \nu_2$, where
354: $\Delta \nu_2 > \Delta \nu_1$.
355: The angular size $\theta_2$ corresponds to this frequency interval.
356: Obviously, $\theta_2 > \theta_1$ in accordance with equation (5).
357: In this case we will see at the difference map the fluctuations with
358: size smaller than $\theta_2$ and the corresponding angular power spectrum
359: will have the break at $l_2\approx 1/\theta_2 <l_1$.
360:
361: In Figs.~1 and 2 the illustration of the differential method is given.
362: In the upper row four maps (1, 2, 3 ¨ 4) observed at four different
363: frequencies $\nu_1, \nu_2, \nu_3, \nu_4$ are shown.
364: All the maps have identical angular power spectrum depicted in Fig.~2 by
365: bold dashed line.
366: The maps shown differ from each other in the following way. Spectrum
367: harmonics of maps with number 1 and number $i (i=2,3,4)$ are
368: characterized by identical phases for $l\le l_s(i)$ (in our case
369: $l_s(2)=640, l_s(3)=1152, l_s(4)=1536)$ and different ones for
370: $l>l_s(i)$. Note that the phases are distributed randomly in interval
371: $[0,2\pi]$.
372:
373:
374: In the bottom row of Fig.~1 three difference maps indicated as 1-2, 1-3
375: and 1-4 obtained after subtraction from the map with number 1
376: of maps with numbers 2, 3 and 4 correspondingly.
377: The frequency differences satisfy the following inequality
378: $\nu_1-\nu_2>\nu_1-\nu_3>\nu_1-\nu_4$.
379: Respectively, maximal size of the CMBR fluctuations visible on the
380: remainder maps meets the inequality: $\theta_{1-2}>\theta_{1-3}>\theta_{1-4}$.
381: Qualitatively, the scale of the fluctuations can be estimated directly
382: from the maps shown in Fig.~1. The angular power spectra of difference maps
383: 1-2, 1-3 and 1-4 are depicted in Fig.~2 by solid, dotted and thin dashed
384: lines correspondingly.
385: As seen the break points of the angular power spectra meet the following
386: relation $l_{1-2}<l_{1-3}<l_{1-4}$. Obviously,
387: $l_s(2)=l_{1-2}, l_s(3)=l_{1-3}, l_s(4)=l_{1-4}$.
388:
389: In the example considered above it was assumed that the observed CMBR maps
390: consist only of the primary and secondary CMBR fluctuations.
391: But in practice the CMBR detected maps contain in addition a number of
392: foregrounds such as pixel noise, extragalactic unresolved sources,
393: free-free, synchrotron radiation, Sunyaev-Zel'dovich effect, Galactic
394: dust and so on.
395: The most dangerous foreground for the differential method considered is
396: the pixel noise. Other foregrounds characterized by considerably
397: low value of spectral index as compared with analyzed secondary CMBR
398: fluctuations and are successfully eliminated in resulted remainder maps.
399: Below we demonstrate the simulation results of the differential method
400: in presence of some additive pixel noise in the CMBR maps.
401:
402: For this Gaussian zero mean white noise was added to the CMBR
403: maps 1,2,3 and 4. Signal-to-noise ratio for each map is $\approx 1.44$.
404: Noisy maps 1,2,3 and 4 are shown in the upper row of Fig.~3.
405: The difference maps 1-2, 1-3, 1-4, obtained after lower-frequency spatial
406: filtering with bandpass equal to the CMBR spectrum width, are shown in
407: the lower row of Fig.~3. As seen from the pictures the presence of the
408: noise in low-frequency region of the CMBR power spectrum led to
409: appreciable mess of fine structure of the remainder fluctuations.
410: Therefore in order to obtain the necessary both qualitative and
411: quantitative characteristics the analysis of the angular power spectra
412: of the difference CMBR maps is required.
413: The angular power spectra of difference maps
414: 1-2, 1-3 and 1-4 are depicted in Fig.~4 by solid, dotted and thin dashed
415: lines correspondingly.
416: As seen from the pictures, the angular power spectrum breaks corresponding
417: to boundaries between the noise spectrum and spectrum of CMBR+noise map
418: are quite apparent even without preliminary processing of the detected
419: CMBR maps.
420: Obviously, in case of higher level of noise a more complicated preliminary
421: signal processing directed to decreasing the noise in the CMBR spectrum
422: region would be required (Bajkova et al., 2002).
423:
424: Availability of the sufficiently accurate measured value of the CMBR angular
425: power spectrum break $l_c$ allows us to estimate $\theta_c$, what in
426: turn allows to determine the retardation parameter $q_0$, redshift $z$,
427: and type of molecules of cosmological origin in accordance with the
428: method explicitly described in (Dubrovich, 1982).
429:
430: \section{Conclusion}
431:
432: Now we formulate the basic results. Presence of molecular clouds in
433: the early stage of the Universe evolution $(z>100)$ must lead to
434: forming spectral-spatial fluctuations of the CMBR.
435: The amplitude of these fluctuations can achieve $310^{-3}K$ in the case
436: of simple scattering and exceed this value in case of sufficiently
437: powerful early extraction of energy what allows to state a problem
438: about their explicit investigation.
439: As shown in this paper, the use of the correlation analysis (Dubrovich,
440: 1982) or
441: considered here more reliable differential method principally allows us
442: both to detect the difference of the primary molecules and determine the
443: type of each of them. Moreover the determination of the interval of $z$,
444: in which these molecules exist as well as parameter $\Omega_m$ is possible.
445:
446: \begin{figure}
447:
448: \centerline{
449: \psfig{figure=fsr1.eps,width=50mm}
450: \hspace{-10mm}
451: \psfig{figure=fsr2.eps,width=50mm}
452: \hspace{-10mm}
453: \psfig{figure=fsr3.eps,width=50mm}
454: \hspace{-10mm}
455: \psfig{figure=fsr4.eps,width=50mm}
456: }
457: {\hskip 11mm 1 \hskip 39mm 2 \hskip 39mm 3 \hskip 39mm 4}
458:
459: \vskip 5mm
460:
461: \centerline{
462: \psfig{figure=d12.eps,width=50mm}
463: \hspace{11.5mm}
464: \psfig{figure=d13.eps,width=50mm}
465: \hspace{11.5mm}
466: \psfig{figure=d14.eps,width=50mm}
467: }
468: {\hskip 9mm 1-2 \hskip 57mm 1-3 \hskip 58mm 1-4}
469: %\caption{}
470: \vskip 5mm
471: \centerline{\bf Fig.~1}
472: \end{figure}
473:
474:
475: \begin{figure}
476:
477: \centerline{
478: \psfig{figure=pas_1.ps,angle=90,width=150mm}
479: }
480: %\caption{}
481: \medskip
482: \centerline{\bf Fig.~2}
483: \end{figure}
484:
485: \begin{figure}
486:
487: \centerline{
488: \psfig{figure=fsr_1.eps,width=50mm}
489: \hspace{-10mm}
490: \psfig{figure=fsr_2.eps,width=50mm}
491: \hspace{-10mm}
492: \psfig{figure=fsr_3.eps,width=50mm}
493: \hspace{-10mm}
494: \psfig{figure=fsr_4.eps,width=50mm}
495: }
496: {\hskip 11mm 1 \hskip 39mm 2 \hskip 39mm 3 \hskip 39mm 4}
497: \vskip 5mm
498:
499: \centerline{
500: \psfig{figure=d_fil12.eps,width=50mm}
501: \hspace{11.5mm}
502: \psfig{figure=d_fil13.eps,width=50mm}
503: \hspace{11.5mm}
504: \psfig{figure=d_fil14.eps,width=50mm}
505: }
506: {\hskip 9mm 1-2 \hskip 57mm 1-3 \hskip 58mm 1-4}
507: %\caption{}
508: \vskip 5mm
509: \centerline{\bf Fig.~3}
510: \end{figure}
511:
512: \begin{figure}
513:
514: \centerline{
515: \psfig{figure=pas_2.ps,angle=90,width=150mm}
516: }
517: %\caption{}
518: \medskip
519: \centerline{\bf Fig.~4}
520: \end{figure}
521:
522: \section{Acknowledgements}
523:
524: We would like to thank Yu. Parijskij for interest to
525: this work.
526:
527: \noindent This work was partially supported in frame of State Contract
528: N 40.022.1.1.1106.
529:
530: \section{References}
531:
532: \noindent Bajkova, A.T., 2002, astro-ph/0205112.
533:
534: \noindent Bennett, D.P., Sun Hong Rhie, 1993, ApJ, {\bf 406}, L7--L10.
535:
536: \noindent Dubrovich, V.K., 1977, Astron. Letters, {\bf 3}, 243 (in
537: Russian).
538:
539: \noindent Dubrovich, V.K., 1982, Izvestiya SAO, {\bf 15}, 21 (in Russian).
540:
541: \noindent Dubrovich, V.K., 1994, A\&A Tr., {\bf 5}, 57.
542:
543: \noindent Dubrovich, V.K., Lipovka, A.A., 1995, A\&A, {\bf 296}, 301.
544:
545: \noindent Dubrovich, V.K., 1997, A\&A, {\bf 324}, 27.
546:
547: \noindent Dubrovich, V.K., 1997a, Pross. HSRA, 22-26 Jan. 1995,
548: Jodrel Bank, CUP, 189.
549:
550: \noindent Dubrovich, V.K., 1999, Gravitation and Cosmology, {\bf 5}, 171.
551:
552: \noindent Dwek, E. et al., 1998, astro-ph/9806129.
553:
554: \noindent Haarsma, D.B., and Partridge, R.B., 1998, ApJ, {\bf 503}, L5.
555:
556: \noindent Hanany, S. et al., 2000, Astrophys. J., {\bf 545}, L5.
557:
558: \noindent Hauser, M.G. et al., 1998, astro-ph/9806167.
559:
560: \noindent Lange, A.E. et al, 2001, Phys. Rev., D63, 042001.
561:
562: \noindent Lepp, S., Shull, J. M., 1984, ApJ, {\bf 280}, 465.
563:
564: \noindent Maoli, R. et al., 1996, ApJ., {\bf 457}, 1.
565:
566: \noindent Palla, F. et al., 1995, ApJ, {\bf 451}, 44.
567:
568: \noindent Puy, D. et al., 1993, A\&A, {\bf 267}, 337.
569:
570: \noindent Sahni, V., Starobinsky, A., 1999, astro-ph/9904398.
571:
572: \noindent Silk, J., 1994, astro-ph/9405072.
573:
574: \noindent Stancil, P.D., Lepp, S., and Dalgarno, A., 1996, ApJ, {\bf 458}, 401S.
575:
576: \noindent Vishniac, E.T., 1987, ApJ, {\bf 322}, 597.
577:
578: \noindent Zel'dovich, Ya.B., 1978, Sov.Astron.Let., {\bf 4}, 165.
579:
580: \end{document}
581:
582: