1:
2: \documentclass[apjl]{emulateapj}
3: %\documentclass[manuscript]{aastex}
4:
5: \usepackage{apjfonts}
6:
7: \usepackage{amsmath}
8: \usepackage{natbib}
9:
10: \usepackage{epsfig}
11:
12: \newcommand{\vdag}{(v)^\dagger}
13: \newcommand{\Lya}{Ly$\alpha$\ }
14: \newcommand{\lya}{Ly$\alpha$\ }
15: \newcommand{\erfc}{\ensuremath{{\rm erfc}}}
16: \newcommand{\psec}{\ensuremath{{\rm s}^{-1}}}
17: \newcommand{\yr}{\ensuremath{{\rm yr}}}
18: \newcommand{\cm}{\ensuremath{{\rm cm}}}
19: \newcommand{\kpc}{\ensuremath{{\rm kpc}}}
20: \newcommand{\Mpc}{\ensuremath{{\rm Mpc}}}
21: \newcommand{\mK}{\ensuremath{{\rm mK}}}
22: \newcommand{\GeV}{\ensuremath{{\rm GeV}}}
23: \newcommand{\TeV}{\ensuremath{{\rm TeV}}}
24: \newcommand{\erg}{\ensuremath{{\rm erg}}}
25: \newcommand{\Hz}{\ensuremath{\, {\rm Hz}}}
26: \newcommand{\MHz}{\ensuremath{\, {\rm MHz}}}
27: \newcommand{\GHz}{\ensuremath{\, {\rm GHz}}}
28: \newcommand{\muGs}{\ensuremath{\, \mu{\rm Gs}}}
29: \newcommand{\sr}{\ensuremath{\, {\rm sr}}}
30: \newcommand{\msun}{\ensuremath{{M_\sun}}}
31: \newcommand{\etal}{{\it et al.}}
32:
33: \begin{document}
34:
35:
36: \title{Angular power spectrum of the galactic synchrotron radiation}
37:
38: \author{Xuelei Chen}
39: \affil{The Kavli Institute for Theoretical Physics, UCSB, Santa
40: Barbara, CA 93106, U.S.A.}
41:
42:
43:
44: \begin{abstract}
45: We calculate the angular power spectrum of the galactic synchrotron
46: radiation induced by the small scale fluctuations of the magnetic field
47: and the cosmic ray electron density.
48: Using the observed interstellar magnetic field
49: spectrum, which is consistent with the Komolgorov turbulence model
50: at the relevant scales, we find that $C_l \propto
51: l^{-3.7}$. We estimate the cosmic ray
52: electron density fluctuation spectrum with
53: an injection-diffusion model, the shape of the angular power spectrum
54: in this model depends on the correlations between the injection
55: sources. For Poisson distribution of sources, $C_l \propto l^{-4}$.
56: We discuss the implications for the interpretation of cosmic
57: microwave background (CMB) data and the impact on future 21 cm
58: tomography experiments.
59: \end{abstract}
60:
61: \keywords{ cosmic microwave background ---
62: radiation mechanism: non-thermal --- radio continuum: galaxies
63: ---ISM: magnetic fields --- cosmic rays}
64:
65: \section{Introduction}
66:
67: The radio sky at $\nu < 1 \GHz$ is dominated by galactic synchrotron
68: emission. It is believed to be produced by cosmic ray electrons propagating
69: in the magnetic field of the Galaxy \citep{GS69}.
70: The galactic synchrotron emission is an important foreground
71: for the cosmic microwave background (CMB)
72: experiments \citep{S99}. For the upcoming high redshift
73: 21cm tomography experiments \citep*{MMR97,TMMR00,CM04,CM03,GS03,
74: F04a,ISMS03,LZ04}, such as
75: PAST \citep{PWP04}, LOFAR\footnote{http://www.lofar.org}, and
76: SKA\footnote{http://www.skatelescope.org}, it poses a major challenge.
77:
78:
79:
80:
81: The global galactic synchrotron emission spectrum from 0.3 MHz to 408
82: MHz can be fit with a two component disk model of
83: the galaxy \citep*{KWL04a,KWL04b}.
84: Analysis of the existing radio surveys at 408 MHz, 1.42 GHz, and 2.326
85: GHz \citep*{H82,R82,RR86,RR88,JBN98} shows that the synchrotron emission
86: has a spectral index $\beta \approx 2.7$ \citep{P03}, which is in general
87: agreement with the CMB result \citep{B03}. The real space
88: distribution of synchrotron emissivity over the galactic disk,
89: taking into account spiral arms, were derived from the 408 MHz whole sky
90: map using refolding techniques \citep*{P81a,P81b,BKB85}.
91: In the Fourier space, the angular power spectrum of the galactic
92: synchrotron radiation can be modeled reasonably well
93: as a simple scaling relation:
94: $C_l \propto l^{-\eta} $ \citep*{TE96,TEHO00,G01,G02}, with $\eta = 2.4-3$
95: down to $l \sim 900$. Recently, the WMAP team obtained a shallower
96: spectrum of $\eta \sim 2$ down to $l \sim 200$. These are extrapolated
97: to higher $l$ in recent studies of the 21 cm foreground \citep*{DMCM04,SCK04}.
98:
99:
100: In the present study we investigate the galactic synchrotron emission from a
101: different perspective. We calculate the synchrotron
102: angular power spectrum induced by the variation of the magnetic field
103: and fluctuations in the cosmic ray electron density.
104: Our primary objective is to achieve a {\it physical} understanding of
105: the origin of the synchrotron emission anisotropy. In
106: particular, we would like to ask which of these two mechanisms
107: is responsible in producing the observed anisotropy?
108: This will help us to assess the validity of the hypothesis
109: adopted in the empirical analyses of CMB and 21 cm observation, e.g.,
110: will the power law form of the angular power spectrum hold
111: down to small scales relevant for the pre-reionization 21 cm observation?
112: At the same time, we may also gain useful knowledge about the galactic
113: distribution of the cosmic ray electrons and the magnetic field.
114: Similar {\it physical} modeling have been performed for a number of
115: other foregrounds, e.g. free-free \citep{OM03,CF04} and intergalactic
116: shocks \citep{KWL04a,KWL04b}.
117:
118: The fluctuation power spectrum of the galactic magnetic field has been
119: measured \citep*{MS96,hfm04}, and is consistent with being produced by
120: a turbulent interstellar medium (ISM) described by the
121: Komolgorov scaling model \citep{K41,ES04}.
122: With this we can calculate directly the anisotropy power spectrum
123: induced by the magnetic field variation. The distribution of the
124: cosmic ray electrons is less well known, but
125: there is a broadly accepted picture of the cosmic electrons being produced in
126: supernovae remnants (SNR), which then diffuses through the whole galaxy and
127: be confined in a volume greater than the galactic disk \citep{GS64}.
128: We can calculate the cosmic ray density distribution with this model.
129:
130:
131: \section{Models}
132:
133: Let us consider the angular power spectrum obtained over a small patch of
134: ``blank'' sky field at high galactic latitude.
135: In the optically thin case, the observed radiation intensity
136: is simply an integral of the
137: emissivity along the line of sight. At very low frequencies, synchrotron
138: self-absorption and plasma absorption are important. For $70 \MHz< \nu <
139: 200 \MHz$, corresponding to the
140: redshift range of 6-20 which is of interest to the study of the
141: reionization process, these can be neglected,
142: \begin{equation}
143: I(\hat{n},\nu) = \int dr ~ \varepsilon(\hat{n}, r, \nu),
144: \end{equation}
145: where $\epsilon$ is the volume emissivity.
146: The emissivity may have both spatial and frequency variations. If the two
147: are separable, we may write
148: \begin{equation}
149: \varepsilon_{\nu}({\bf x})= g(\nu) \psi({\bf x}), \qquad
150: \tilde\varepsilon_{\nu}({\bf k})=g(\nu) \tilde \psi({\bf k});
151: \end{equation}
152: where the tilde denotes Fourier transform,
153: $\psi({\bf x}) = \int \frac{d^3 k}{(2\pi)^3} e^{i{\bf k}\cdot {\bf x}}
154: \tilde{\psi}({\bf k})$.
155: The emissivity power spectrum is then
156: \begin{equation}
157: P_{\varepsilon}(\nu_1, \nu_2, \mathbf{k}) = g(\nu_1) g(\nu_2) P_{\psi}(\mathbf{k}),
158: \end{equation}
159: Let us consider the case $\nu_1=\nu_2=\nu$.
160: At small scale, using the Limber
161: approximation \citep{K92,ZFH04}, the angular power spectrum is given by
162: \begin{equation}
163: \label{eq:cl}
164: C_l^T (\nu)= \frac{c^4 g^2(\nu)}{4\nu^4 k_B^2}
165: \int \frac{dr}{r^2} P_{\psi}(\frac{l}{r}).
166: \end{equation}
167: If the emissivity has a power law spectrum, $P_{\psi}(k) \propto k^{\eta}$,
168: then $C_l \propto l^{\eta}$. The integral is up to some cutoff point,
169: at which the emissivity drops to 0.
170:
171: We now consider the emissivity of synchrotron radiation.
172: If the energy distribution of cosmic ray
173: electrons $f(E)$ at each point is approxiamated as a simple power law with
174: $f(E) = C E^{-p}$,
175: then
176: \begin{equation}
177: \varepsilon(\nu) =\frac{\sqrt{3} e^3}{m_e c^2} C B_{\perp} \alpha_{syn}(p)
178: \left(\frac{2\pi m_e c\nu}{3eB_{\perp}}\right)^{-(p-1)/2}
179: \end{equation}
180: where $\alpha_{syn}(p)=\frac{1}{p+1}\Gamma[\frac{3p-1}{12}]
181: \Gamma[\frac{3p+19}{12}]$.
182: We see that for power law distribution with $f(E) \propto E^{-p}$,
183: $I(\nu) \propto \nu^{\alpha}$, and $T(\nu) \propto \nu^{\beta}$, where
184: $\alpha=-(p-1)/2$, and $\beta=\alpha-2$. A
185: power law is actually a good approximation to the real
186: distribution function of cosmic ray electrons.
187: In this approximation, variation of emissivity can be
188: induced by varying the magnetic field $B$, spectral index $p$, and
189: cosmic ray electron density normalization $C$.
190: Here we shall fix $p$ and consider the variation of $B$ and $C$.
191:
192:
193: The galactic magnetic field is typically a few $\muGs$, for this
194: magnetic field the emission at $70-200 \MHz$ is produced primarily
195: by cosmic electrons with $E \sim 0.1 ~\GeV - 10 ~\GeV$.
196: The local cosmic ray electron density
197: can be measured directly. There is some uncertainty in the
198: normalization, and corrections have to be made for solar modulation.
199: A recent compilation of measurements yields \citep{CB04}, in our units,
200: $C=1.7 \times 10^{-11} \cm^{-3} \GeV^{-1}$,
201: and $p=3.44$, at the energy of a few GeV.
202: The spectral index $p=3.4$ yields $\alpha= -1.2$, and $\beta=
203: -3.2$. For comparison, the WMAP measurement \citep{B03}
204: indicates that towards star
205: forming regions, $\beta=-2.5$, and towards the halo,
206: $\beta=-3.0$. Radiative loss of electron energy provides a natural
207: explanation to the steeping of the electron spectrum away from
208: star-forming region. Below 3 GeV, the electron spectrum becomes
209: flatter, probably because the primary energy loss mechanism changes from
210: radiation to ionization.
211: The locally measured electron density may not
212: represent the average, nevertheless we will use it as a trial value.
213: We will take $p=3$, which gives a slightly better fit to the radio
214: data than the steep $p=3.4$ value.
215: We also assume a disk scale height of 1 kpc (corresponding to the
216: thick disk in the \citealt{KWL04a} model), and a smooth magnetic field
217: of $4 \muGs$. With this set of parameters, the integrated sky
218: brightness temperature at 408 MHz is 20 K, which reproduces the
219: observed value at high galactic latitude \citep{H82}.
220:
221:
222: Let us consider first the magnetic field with $C$ fixed. Now
223: \begin{equation}
224: \psi= b(x) = B_{\perp}^{(p+1)/2}(x).
225: \end{equation}
226: If there is a large smooth global magnetic field component which
227: varies only on large scale and
228: fluctuations around it are small, then we can write $B=B_0+\delta
229: B(x)$, and
230: \begin{equation}
231: B_{\perp}^{(p+1)/2}=B_{0\perp}^{(p+1)/2}\left(1+\frac{p+1}{2}
232: \frac{\delta B_{\perp}(x)}{B_{0\perp}} + ...\right)
233: \end{equation}
234: then
235: \begin{equation}
236: P_{\psi}(k) = \frac{(p+1)^2}{6} C^2 B_{0\perp}^{p-1} P_{\delta B} (k) .
237: \end{equation}
238:
239:
240: The interstellar medium is turbulent \citep{ES04}. Komolgorov derived
241: a scaling relation for scale-invariant turbulence \citep{K41},
242: with a power spectrum of the form $k^{-11/3}$.
243: The magnetic field fluctuation spectrum can be determined by
244: energy equipartition. The predicted magnetic field fluctuation
245: spectrum is confirmed on the
246: scales of 0.01 pc -- 100 pc by observation of the Faraday rotation
247: of extragalactic sources \citep{MS96}. At larger scales and on the
248: galactic disk, the magnetic field spectrum is flatter, probably
249: because at these scales the motion is dominated by
250: by two-dimensional structure (vortices). If we join the
251: small scale and the large scale observations \citep{hfm04}, we obtain
252: \begin{equation}
253: \label{eq:Bspec}
254: E(k) = \left\{ \begin{array}{ll}
255: 2.03 \times 10^{-11} ~
256: k_{\kpc^{-1}}^{-5/3}
257: ~\erg~ \cm^{-3}~ \kpc,& k_{\kpc^{-1}}
258: > 1.57 \times 10^3 \\
259: 1.34 \times 10^{-12} ~
260: k_{\kpc^{-1}}^{-0.37}
261: \erg~ \cm^{-3}~ \kpc, & k_{\kpc^{-1}} < 6.28 \\
262: \end{array}\right.
263: \end{equation}
264: This is plotted in Fig.~\ref{fig:Bspec} (Note that our definition of k differs from \cite{hfm04}
265: by a factor of $2\pi$). The energy is related to the
266: power spectrum by
267: \begin{equation}
268: P_B(k) = 2 E_B(k)/k^2.
269: \end{equation}
270: Other scaling models have also been suggested. For example, some MHD
271: turbulence have $E \sim k^{-3/2}$ instead of $k^{-5/3}$ \citep{K65}.
272: However, this would produce an angular power spectrum not very
273: different from the Komolgorov one.
274:
275: \begin{figure}
276: \plotone{Fig1.eps}
277: \caption{\label{fig:Bspec} The magnetic field energy spectrum as extrapolated
278: from \cite{MS96} and \cite{hfm04}. The dotted part is interpolation.}
279: \end{figure}
280:
281:
282: We can then carry out the calculation, with
283: the magnetic field spectrum given in Eq.~(\ref{eq:Bspec}) and
284: the locally measured electron density. However,
285: even though we keep $C$ fixed to investigate the variation induced by
286: the magnetic field, in reality it must decrease as we move away from
287: the galactic disk. We can approximate this effect by taking the power
288: spectrum $P_{\psi}(k)$ also as an explicit function of $r$, i.e.
289: $P_{\psi}(k,r) \sim P(k) e^{-2r/r_0}$, where
290: $r_0$ is the scale height of the halo in which the cosmic ray
291: electrons are confined. It turns out that this damping factor only
292: changes the result by a small factor, because the large $r$
293: contribution is already suppressed by the $P \sim (l/r)^{-5/3}$ factor.
294: For the small scale that we are mostly interested in, Komolgorov
295: spectrum applies, $P_B(k) \sim k^{-11/3}$, and
296: $C_l \sim l^{-3.7}$.
297:
298: The result of our calculation for
299: $\nu=150 \MHz$ (21 cm line at $z \sim 9$),
300: and $ 23 \GHz$ are plotted in Fig.~\ref{fig:cl} as solid lines, along
301: with the WMAP K band data (centered at 23 GHz).
302: Remarkably, at small $l$, the anisotropy power amplitude is
303: of the same order of magnitude as the WMAP data (At such small $l$,
304: the Limber approximation may not be very good, nevertheless the true
305: result will be of the same order of magnitude). This means that the
306: magnetic field fluctuation may play a role in the formation of the
307: observed synchrotron anisotropy. However, it is clear that compared
308: with the data $C_l$ drops too fast as $l$ increases.
309:
310:
311: \begin{figure}
312: \plotone{Fig2.eps}
313: \caption{\label{fig:cl} The synchrotron power spectrum for
314: $\nu=150\MHz$ (upper curves) and $\nu=20\GHz$ (lower curves).
315: The solid line marks the anisotropy induced by the
316: fluctuation of magnetic field, the dashed line for that induced by the
317: fluctuation of cosmic ray
318: electron density, the points are WMAP K band (23 GHz) data. }
319: \end{figure}
320:
321:
322:
323: If the smooth magnetic field component is absent and the
324: fluctuating field dominates, then
325: \begin{equation}
326: P_{\psi} = \langle B_{\perp}^{p+1} \rangle (k) =
327: \langle B^{p+1} \rangle (k) c(p)
328: \end{equation}
329: where
330: $c(p)= \frac{1}{2}\int_0^{\pi} d\theta ~ \sin^{p+2}\theta=
331: \sqrt{\pi}~ \Gamma[(3+p)/2]/\Gamma[(2+p/2)]$.
332: In this case, the result is uncertain,
333: because it depends on the higher order
334: correlation $\langle P_{B^{(p+1)/2}}(k) \rangle $.
335: If we make the gaussian-like {\it ansatz} $\langle B^{p+1}(k)\rangle
336: \sim (\langle B^2(k)\rangle)^{(p+1)/2} $, and $P_B(k) \sim k^{-\eta_B}$,
337: then the result is
338: $C_l \propto l^{-(p+1)\eta_B/2}$. For
339: the Komolgorov spectrum $\eta_B = -11/3$, $C_l \sim l^{-7.3}$,
340: which is extremely steep and can be neglected entirely. However, this
341: assumption maybe incorrect.
342:
343:
344:
345:
346:
347: Now we consider the variation of the cosmic ray electron density. This
348: is not known from observation. To make an estimate of the fluctuation,
349: we consider an injection-diffusion model \citep{GS64}, in which
350: the cosmic ray electrons are produced (injected) in point sources,
351: then diffuse out, until eventually losing all their energy or
352: escaping the confinement volume. The density of electrons at a point in
353: space is then dependent on the distance to nearby sources.
354: Neglecting the momentum space diffusion, the density
355: normalization constant at position $\mathbf{r}$, time $t$ satisfies
356: the diffusion equation
357: \begin{equation}
358: \frac{\partial C(\mathbf{r},t)}{\partial t} = D \nabla^2
359: C(\mathbf{r},t) + q(\mathbf{r},t) - \frac{C(\mathbf{r},t)}{\tau}
360: \end{equation}
361: where $q(\mathbf{r},t)$ is the cosmic ray electron injection rate at
362: $\mathbf{r}$, and $D$ is the diffusion coefficient,
363: which is constant below 5 GeV,
364: and $D \approx 2\times 10^{28} (E/5\GeV)^{0.6}$ at $E > 5 \GeV$
365: \citep{KKYN04}. The last term in the equation represents loss of
366: electrons by radiation, ionization, etc., or by escaping the
367: confinement volume, with $\tau$ the loss time scale. For radiative
368: losses \citep{CB04},
369: \begin{equation}
370: \tau \approx 2.1 \times 10^5 (E/\TeV)^{-1} \yr.
371: \end{equation}
372: We note that in the injection-diffusion model, the scale height of the
373: cosmic ray halo is a few times of
374: $\times \sqrt{D \tau}$, with the above values we have
375: $\sqrt{D\tau} \sim 0.3 \kpc$.
376: Although very crude, our model is self-consistent.
377: In Fourier space,
378: \begin{equation}
379: \frac{\partial \tilde{C}(k)}{\partial t} + (D k^2+\frac{1}{\tau})
380: \tilde{C}(k) = \tilde{q}(k)
381: \end{equation}
382: The steady state solution is
383: \begin{equation}
384: \tilde{C}(k) = \frac{\tilde{q}(k)}{D k^2 +\frac{1}{\tau}}.
385: \end{equation}
386: The power spectrum is then
387: \begin{equation}
388: P_C(k) = \frac{P_q(k)}{(D k^2 +\frac{1}{\tau})^2}.
389: \end{equation}
390:
391:
392:
393: Supernovae remnants are most likely the primary source for these cosmic
394: ray electrons. The injection function is then $q = N_e \kappa_{SN}$,
395: where $N_e$ is the number of
396: cosmic ray electrons produced in one supernova,
397: and $\kappa_{SN}$ is the number of supernovae explosions per unit
398: volume per unit time. If the distribution of supernovae is Poisson,
399: with $\langle \kappa_{SN} \rangle =1/(V t_{SN})$,
400: where $V$ is the volume in which
401: one supernova explode per average interval $t_{SN}$, then
402: \begin{equation}
403: P_C(k) = \frac{N_e^2 V}{(D k^2 +\frac{1}{\tau})^2 t_{SN}^2 V^2}
404: \end{equation}
405:
406: When $k \to 0$, we have $P_C(k) = C_0^2 V$, where $C_0$ is the average
407: of $C$. From this we obtain
408: \begin{equation}
409: \frac{N_e}{V} = C_0 \frac{t_{SN}}{\tau}
410: \end{equation}
411: As a reality check, we take $E \sim 1 \GeV$,
412: a stellar disk with radius of 15 kpc,
413: and a scale height of 300 pc, and also assume that in this volume the
414: average interval of supernovae explosion is 50 years, then we find
415: $N_e \sim 10^{48} $, which requires an energy of $10^{45} \erg$,
416: which is a small fraction of the total energy of a supernova, hence
417: supernovae do have sufficient energy to generate these cosmic ray electrons.
418:
419: Using this relation we finally obtain
420: \begin{equation}
421: P_C = \frac{C_0^2}{ (D k^2 \tau +1)^2} V,
422: \end{equation}
423: Note that when written in this form,
424: the result does not depend on $t_{SN}$.
425: The resulting $C_l$ is plotted as dashed curves in Fig.~\ref{fig:cl}.
426: At large angle, the predicted $C_l$ in this model is comparable to the
427: case of magnetic field induced fluctuation, and agrees with what is
428: observed at the order of the magnitude level. However, at large $l$,
429: $C_l \sim l^{-4}$, again the spectrum is too steep compared with
430: observation. Although the spectrum flattens at small $k$, this
431: happened only at scales comparable to the disk scale height, thus
432: affecting only large angle ($l \sim$ a few).
433:
434: In the above we have assumed a Poisson distribution of supernovae
435: remnant. If they are correlated, with
436: $P_{SN}(k) \propto A k^{\gamma}$, then
437: $C_l \propto l^{\gamma-4} $.
438: Supernovae may well be correlated, as their rate should be
439: proportional to the star formation rate, which in turn depends on the
440: density. However, the Komolgorov model of turbulence suggests
441: $\gamma<0$. The observation of flocculent star light distribution
442: in nearby spiral galaxies seem to confirm this expectation,
443: which has a power law of $P_{k} \sim k^{-1}$ at the 100 pc
444: scale \citep*{EEL03,E03}. If so, then the correlation of SNR may not
445: help us. However, more observations are needed to address this
446: issue.
447:
448:
449:
450:
451: \section{Conclusion}
452: We have calculated angular power spectrum of the galactic synchrotron
453: radiation induced by the variation of the magnetic field and cosmic
454: ray electron density. We found that at low $l$, the amplitudes of the
455: anisotropy power produced by both mechanisms are comparable to the
456: observed value. This indicates that these physical mechanisms
457: {\it are} relevant to the production of the observed foreground anisotropy.
458: However, neither of these two mechanisms can produce the
459: $C_l \sim l^{-2}$ power spectrum observed by WMAP \citep{B03}, and are also
460: steeper than the older value of $\eta \sim 2.4-3 $.
461: The magnetic field model induce power spectrum of the form
462: $C_l \sim l^{-3.7}$, while for the electron density fluctuation
463: it is $C_l \sim l^{-4}$ if the spatial distribution of the SNR is
464: Poisson. This may be remedied if the spatial distribution of supernova
465: remnants has the form $P_{SN}(k) \sim k^2$. However, observations seem
466: to indicate $P_{SN}(k) \sim k^{-1}$.
467:
468: We made a number of simplifications in our calculation.
469: We did not consider the detailed distribution of global
470: magnetic field and cosmic ray sources in the Galaxy, nor do we
471: consider accelerations outside SNR. It is unlikely that inclusion of
472: any of these details would change our qualitative conclusion.
473: Perhaps more important is our assumption of a universal energy spectrum of
474: the electrons. In reality, the spectral index varies from place to
475: place, and this may induce additional anisotropy. To model this, we
476: need to include momentum space diffusion. We plan to address
477: these issues in subsequent studies.
478:
479: If the galactic synchrotron emission does have a steep
480: angular power spectrum as suggested here, how could we
481: reconcile this with the shallower power spectrum reported by the WMAP
482: team? One intriguing possibility is that there may be
483: another type of foreground, which has a frequency dependence similar
484: to the synchrotron radiation at the relevant bands, and was mistaken as
485: synchrotron radiation. Recently, several groups of researchers have
486: suggested that some of the foreground identified as synchrotron by the
487: WMAP team may actually be spinning dust \citep{L03,F03,dO04}.
488: This has a peak at $10 \GHz - 20 \GHz$, which would not affect the
489: 21 cm measurements. There might also be other foregrounds, e.g. of
490: extragalactic origin, which contributes to the radio
491: survey. Alternatively, there may be other unknown mechanisms which
492: produce the small scale anisotropy in the galactic
493: synchrotron radiation. Further investigations are needed to
494: identify what is responsible for producing the small scale anisotropy
495: power. For the two known physical mechanisms discussed here, the
496: anisotropy power at small scales is much smaller than derived from simple
497: extrapolation.
498:
499:
500:
501: \acknowledgments
502: I thank J. L. Han, E. Scannapieco, S. Furlanneto, D. Casadei and
503: S. P. Oh for suggestions and discussions.
504: This work is supported by the NSF grant PHY99-07949.
505:
506:
507:
508:
509: \begin{thebibliography}{}
510:
511: \bibitem[Bennett \etal(2003)]{B03} Bennett, C. L. et al., 2003, \apjs,
512: 148, 97.
513:
514: \bibitem[Beuermann \etal(1985)Beuermann, Kanbach,\& Bekhuijsen]{BKB85}
515: Beuermann, K., Kanbach, G., \& Bekhuijsen, E.M., 1985, \aap, 153, 17.
516:
517: \bibitem[Casadei \& Bindi(2004)]{CB04} Casadei, D. \& Bindi, V., 2004,
518: \apj, 612, 262.
519:
520: \bibitem[Chen \& Miralda-Escud\'{e}(2004)]{CM04} Chen, X.
521: \& Miralda-Escud\'{e}, J., 2004, \apj, 602, 1.
522:
523: \bibitem[Ciardi \& Madau(2003)]{CM03} Ciardi, B. \& Madau, P.,
524: astro-ph/0303249.
525:
526: \bibitem[Cooray \& Furlanetto(2004)]{CF04} Cooray, A. \& Furlanetto,
527: S. R., 2004, astro-ph/0402239
528:
529: \bibitem[de Oliveira-Costa \etal(2004)]{dO04} de Oliveira-Costa, A.,
530: Tegmark, M., Davies, R. D., Gutierrez, C. M., Lasenby, A. N.,
531: Rebolo, R., \& Watson, R. A., 2004, \apjl, 606, 89.
532:
533: \bibitem[Di Matteo \etal(2004)Di Matteo, Ciardi, \& Miniati]{DMCM04}
534: Di Matteo, T., Ciardi, B., \& Miniati, F. 2004, astro-ph/0402332
535:
536: \bibitem[Elmegreen \etal(2003)Elmegreen, Elmegreem, \& Leitner]{EEL03}
537: Elmegreen, B. G., Elmegreen, D. M., \& Leitner, S. N., 2003, \apj, 590, 271.
538:
539: \bibitem[Elmegreen \etal(2003)]{E03}
540: Elmegreen, B. G., Leitner, S. N., Elmegreen, D. M., \& Cuillandre, J-C.,
541: 2003, \apj, 593, 333.
542:
543:
544: \bibitem[Elmegreen \& Scalo(2004)]{ES04} Elmegreen, B. G. \& Scalo,
545: J.,2004, astro-ph/0404451.
546:
547: \bibitem[Finkbeiner(2003)]{F03} Finkbeiner, D. P., 2003, astro-ph/0311547
548:
549:
550: \bibitem[Furlanetto \etal(2004a)Furlanetto, Sokasian, \& Hernquist]{F04a}
551: Furlanetto,S.R., Sokasian, A., \& Hernquist, L., 2004a, \mnras, 347, 187
552:
553: \bibitem[Furlanetto \etal(2004b)Furlanetto, Zaldarriaga, \& Hernquist]{F04b}
554: Furlanetto,S.R., Zaldarriaga, M., \& Hernquist, L., 2004b, astro-ph/0404112
555:
556: \bibitem[Giardino \etal(2001)]{G01}
557: Giardino, G., Banday, A. J., Foalba, P., Gorski, K. M., , Jonas,
558: J. L., O'Mullane, W., Tauber, J., 2001, \aap, 371, 708.
559:
560: \bibitem[Giardino \etal(2002)]{G02}
561: Giardino, G., Banday, A. J., Gorski, K. M., Bennett, K., Jonas, J. L.,
562: Tauber, J., 2002, \aap, 387, 82.
563:
564: \bibitem[Ginzburg \& Syrovatskii(1964)]{GS64} Ginzburg, V. L. \&
565: Syrovatskii, S. I., 1964, {\it The Origin of cosmic rays}, Pergamon
566: Press, Oxford.
567:
568:
569: \bibitem[Ginzburg \& Syrovatskii(1969)]{GS69} Ginzburg, V. L. \&
570: Syrovatskii, S. I., 1969, \araa, 7, 375.
571:
572:
573:
574: \bibitem[Gnedin \& Shaver(2003)]{GS03} Gnedin, N. Y. \& Shaver, P.A.,
575: astro-ph/0312005
576:
577:
578: \bibitem[Han \etal(2004)Han, Ferriere, \& Manchester]{hfm04}
579: Han, J. L., Ferriere, K., Manchester, R. N., astro-ph/0404221
580:
581:
582: \bibitem[Haslam \etal(1982)]{H82} Haslam, C. G. T., Salter, C. J.,
583: Stoffel, H., Wilson, W. E., 1982, A\&AS, 47, 1.
584:
585:
586: \bibitem[Iliev \etal(2003)]{ISMS03} Iliev, I. T., Scannapieco, E.,
587: Martel, H., Shapiro, P. R., 2003, \mnras, 341,8.
588:
589: \bibitem[Jonas \etal(1998)Jonas, Baart, \& Nicolson]{JBN98} Jonas,
590: J.L., Baart, E.E., Nicolson, G.D., \mnras, 1998, 297, 977.
591:
592: \bibitem[Kaiser(1992)]{K92} Kaiser, N., 1992, \apj, 388, 272.
593:
594: \bibitem[Keshet \etal(2004a)Keshet, Waxman, \& Loeb]{KWL04a}
595: Keshet, U., Waxman, E., \&
596: Loeb, A., 2004a, astro-ph/0402320
597:
598: \bibitem[Keshet \etal(2004b)Keshet, Waxman, \& Loeb]{KWL04b}
599: Keshet, U., Waxman, E.,
600: \& Loeb, A., 2004b, astro-ph/0407243
601:
602:
603: \bibitem[Kobayashi \etal(2004)]{KKYN04} Kobayashi, T., Komori, Y.,
604: Yoshida, K., Nishimura, J. 2004, \apj, 601, 340.
605:
606: \bibitem[Komolgorov(1941)]{K41} Kolmogorov, A. N., 1941,
607: Proc. Roy. Soc. London Ser. A., 434, 9.
608:
609: \bibitem[Kraichnan(1965)]{K65} Kraichnan, R. H., 1965,
610: Phys. Fluids. 8, 1385.
611:
612: \bibitem[Lagache(2003)]{L03} Lagache, G., 2004, \aap, 405, 813.
613:
614: \bibitem[Loeb \& Zaldarriaga(2004)]{LZ04} Loeb, A. \& Zaldarriaga, M.,
615: 2004, \prl, 92, 212301
616:
617:
618: \bibitem[Madau \etal(1997)Madau, Meiksin, \& Rees]{MMR97} Madau, P., Meiksin, A.,
619: Rees, M. J., 1997, \apj, 475, 429.
620:
621: \bibitem[Minter \& Spangler(1996)]{MS96} Minter, A. H. \& Spangler,
622: S. R., \apj, 458, 194.
623:
624: \bibitem[Oh \& Mack(2003)]{OM03} Oh, S.P., \& Mack, K. J., 2003, \mnras,
625: 346, 871
626:
627: \bibitem[Pen \etal(2004)Pen, Wu, \& Peterson]{PWP04} Pen, U., Wu, X.P.,
628: Peterson, J., astro-ph/0404083.
629:
630: \bibitem[Phillips \etal(1981a)]{P81a} Phillipps, S., Kearsey, S.,
631: Osborne, J. L., Haslam, C. G. T., Stoffel, H., 1981a, \aap, 98, 286.
632:
633:
634: \bibitem[Phillips \etal(1981b)]{P81b} Phillipps, S., Kearsey, S.,
635: Osborne, J. L., Haslam, C. G. T., Stoffel, H., 1981b, \aap, 103, 405.
636:
637:
638:
639:
640: \bibitem[Platania \etal(2003)]{P03} Platania, P., Burigana, C.,
641: Maino, D., Caserini, E., Bersanelli, M., Cappellini, B., Mennella,
642: A., 2003, \aap, 410, 847.
643:
644: \bibitem[Reich(1982)]{R82} Reich, W., 1982, A\&AS, 48, 219.
645:
646: \bibitem[Reich \& Reich(1986)]{RR86} Reich, P. \& Reich, W., 1986, A\&AS, 63, 205.
647:
648: \bibitem[Reich \& Reich(1988)]{RR88} Reich, P. \& Reich, W., 1988, A\&AS, 74, 7.
649:
650:
651: \bibitem[Santos \etal(2004)Santos, Cooray, \& Knox]{SCK04}
652: Santos, M. G., Cooray, A., Knox, L., 2004, astro-ph/0408515
653:
654: \bibitem[Smoot(1999)]{S99} Smoot, G. F., 1999, ``CMB synchrotron
655: foreground'', ASP conf. Ser. 181, 61 (astro-ph/9902201).
656:
657: \bibitem[Tegmark \& Efstathiou(1996)]{TE96} Tegmark, M. \&
658: Efstathiou, G., 1996, \mnras, 281, 1297.
659:
660: \bibitem[Tegmark \etal(2000)Tegmark,Eisenstein,de Oliveira-Costa]{TEHO00}
661: Tegmark, M., Eisenstein, D. J.,
662: Hu, W., de Oliveira-Costa, A., 2000, \apj, 530, 133.
663:
664: \bibitem[Tozzi et al.(2000)]{TMMR00} Tozzi, P., Madau, P. Meiksin, A.,
665: \& Rees, M. J., 2000, \apj, 528, 597.
666:
667: \bibitem[Zaldarriaga \etal(2004)Zaldarriaga, Furlanetto, \& Hernquist]{ZFH04}
668: Zaldarriaga, M., Furlanetto, S.R., \& Hernquist, L., 2004, \apj, 608, 622.
669:
670: \end{thebibliography}
671:
672:
673:
674: \end{document}
675: