1: \documentclass{emulateapj} % ApJ format
2: %\usepackage{apjfonts}
3: %\usepackage{epsf}
4: %\documentclass[12pt,preprint]{aastex}
5:
6: \begin{document}
7:
8: \title{PREGALACTIC LiBeB PRODUCTION BY SUPERNOVA COSMIC RAYS}
9:
10: \author{MOTOHIKO KUSAKABE\altaffilmark{1}}
11: \affil{Department of Astronomy, School of Science, University of
12: Tokyo, Hongo, Bunkyo-ku, Tokyo 113-0033, Japan \\
13: Division of Theoretical Astronomy, National Astronomical
14: Observatory of Japan, Mitaka, Tokyo
15: 181-8588, Japan\\
16: {\tt kusakabe@th.nao.ac.jp}}
17:
18: \altaffiltext{1}{Research Fellow of the Japan Society for the Promotion
19: of Science.}
20:
21: \begin{abstract}
22: I calculate the evolution of Be and B abundances produced by cosmic
23: rays generated by massive stars in the pregalactic phase of the
24: universe. The inputs for calculation, i.e. the star formation rate
25: and the nuclear abundances of cosmic rays, which I assume to be the
26: same as those of the ISM, are taken from the results of a detailed
27: cosmic chemical evolution model with its parameters best fitted from
28: several items of observational information including an early reionization of
29: the IGM by $z\sim 15$. I found that when the $^6$Li plateau abundance
30: observed in metal-poor halo stars originated in the pregalactic
31: cosmological cosmic ray nucleosynthesis, Be and B simultaneously
32: produced with $^6$Li amount to the lowest levels ever detected in
33: metal-poor halo stars. It is desirable to observe Be and B abundances
34: in metal-poor halo stars with [Fe/H]$\leq -3$ in order to elucidate the
35: possibility of early $^6$LiBeB production by pregalactic supernova
36: cosmic ray nucleosynthesis.
37: \end{abstract}
38:
39: \keywords{cosmic rays --- cosmology: theory --- nuclear reactions,
40: nucleosynthesis, abundances --- stars: abundances --- stars: Population
41: II --- supernovae: general}
42:
43: \section{INTRODUCTION}
44:
45: The lithium abundances observed in metal-poor halo stars (MPHSs) show a
46: plateau as a function of metallicity
47: ~\citep{spi1982,rya2000,mel2004,asp2006,bon2007,shi2007} at $^7$Li/H$=1-2$
48: $\times 10^{-10}$. The
49: prediction by the standard big bang nucleosynthesis (BBN) model of
50: $^7$Li abundance which is the main lithium isotope observed in MPHSs, however, indicates
51: a factor of $2-4$ larger value, when the baryon-to-photon ratio deduced
52: from parameter fits to the temperature fluctuations of cosmic microwave
53: background (CMB) radiation measured with Wilkinson Microwave Anisotropy Probe
54: (WMAP)~\citep{spe2003,spe2007} is
55: used. For example, \citet{coc2004} derived $^7$Li/H=$(4.15^{+0.49}_{-0.45})\times 10^{-10}$
56: with the baryon-to-photon ratio $\eta=(6.14\pm 0.25)\times 10^{-10}$.
57: This discrepancy between the observations and the BBN+CMB prediction of
58: $^7$Li abundance is a problem, which indicates some destruction
59: process of $^7$Li. Recently, a complex but consistent theory is
60: suggested by~\citet{pia2006}. In their theory, an extremely high
61: efficiency of engulfment of baryons in a first generation of stars
62: results in a half or one third of destructions of D and Li isotopes.
63: Population II (Pop II) stars are made of a mixture of ejecta of
64: supernovae (SNe) of the
65: first stars and unprocessed material of the primordial composition, and
66: experience a depletion of lithium isotopes in their atmospheres, which
67: are observed. There are two important conditions. The ejecta of SNe of the first stars needs to mix with the pure BBN-composition
68: matter at $2.5\leq$ [Fe/H]\footnote{[A/H]=log(A/H)$-$log(A/H)$_{\sun}$} in order to form the lithium plateau. An
69: infall of the intergalactic medium (IGM) needs to proceed after the formation
70: of Pop II stars in order to be consistent with observations of deuterium
71: abundance.
72:
73: Recent spectroscopic observations of MPHSs also provide abundances of
74: $^6$Li isotope. They indicate a likely primordial plateau abundance,
75: similar to the well known $^7$Li plateau, of
76: $^6$Li/H=$6\times10^{-12}$, which is about 1000 times as large as the
77: BBN prediction. Since the standard Galactic
78: cosmic ray (CR) nucleosynthesis models predict negligible amounts of $^6$Li
79: abundance with respect to the observed plateau level at [Fe/H] $<-2$~\citep[e.g.][]{pra2006}, this plateau causes another problem, which indicates some production process of $^6$Li.
80:
81: Several candidates for early $^6$Li production mechanisms have been suggested. The non-thermal nuclear reactions triggered by the decay of
82: long-lived particles is one possibility of the non-standard
83: process~\citep{jed2000,jed2004a,jed2004b,jed2006,kawasaki2005,kus2006,cum2007}.
84: \citet{pos2006} suggested the exotic nuclear reaction of
85: $^4$He$_X$($d$,$X^-$)$^6$Li to make abundant $^6$Li, where $X^-$ is a negatively charged massive
86: particle assumed to decay in the early universe, and $^4$He$_X$ is the
87: state that has $^4$He bound to $X^-$.
88: \citet{suz2002} suggested an $\alpha+\alpha$ fusion reaction with $\alpha$
89: particles accelerated by hierarchical structure formation shocks, thought
90: to have been operative at the Galaxy formation epoch.
91:
92: As a possibility, \citet*{rol2005} have calculated the $^6$Li
93: production by an initial burst of cosmological cosmic rays (CCRs) to
94: show that this process through $\alpha+\alpha$ fusion can account for the $^6$Li plateau without
95: overproduction of $^7$Li. \citet{rol2006} applied the CCR
96: nucleosynthesis to a well grounded detailed model. They
97: derived the total CCR energy as a function of redshift from a star
98: formation rates (SFRs) in models of cosmic chemical evolution of
99: \citet{dai2006}, which are made to reproduce the observed cosmic SFR, SN
100: II rate, the present fraction of baryons in structures, that in stars,
101: the evolution of the metal content in the interstellar medium (ISM) and IGM, and early
102: reionization of the IGM. As a result they found that the pregalactic
103: production of the $^6$Li in the IGM via Population III (Pop III) stars can account
104: for the $^6$Li plateau without overproduction of $^7$Li.
105:
106: The CR nucleosynthesis also produces Be and B by spallation
107: reactions between CNO nuclei and $p$ and $\alpha$ particles. If this CCR
108: nucleosynthesis scenario of $^6$Li production leads to overproduction
109: of Be and B nuclides against observations in metal-poor stars, it cannot
110: be achieved in the real universe. This study is devoted to checking if the CCR
111: nucleosynthesis as a mechanism of $^6$Li production is consistent with observations of Be and B in
112: metal-poor stars, and finding constraints on conditions of the CCR
113: nucleosynthesis.
114:
115: Abundances of Be and B are observed in MPHSs. A trend of $^9$Be
116: abundance is found such that Be increases linearly as Fe during the
117: course of Galactic evolution~\citep{boe1999}. The first report on
118: observation of beryllium in two very metal-poor stars with the
119: Ultraviolet and Visible Echelle Spectrograph (UVES) mounted on the ESO
120: VLT Kueyen telescope~\citep{pri2000a} said that the
121: trend of beryllium with metallicity keeps decreasing at lower
122: metallicities with no evidence of flattening. \citet{pri2000b} found that the very metal deficient star G 64-12 ([Fe/H]=$-3.3$) has Be of
123: log (Be/H)$=-13.10~\pm$~0.15 dex, which is significantly higher than
124: expected from the previous trend, and claimed that this high [Be/Fe] ratio
125: may suggest a flattening in the beryllium evolutionary trend at the
126: lowest metallicity end or the presence of dispersion at early epochs of
127: the Galactic evolution. On the other hand,~\citet{boe2006} found that G64-37 with [Fe/H]=$-3.2$ has a Be abundance which is
128: consistent with the Be-Fe trend, and suggested that different Be values
129: are indicative of a Be dispersion even at the lowest metallicities.
130:
131: B abundances in metal-poor stars have been estimated by observations
132: with the Goddard High Resolution Spectrograph (GHRS) on board the {\it Hubble
133: Space Telescope} ({\rm HST})~\citep{dun1997,gar1998,pri1999,cun2000}. The boron abundances are also
134: found to show a linear increase with a slope of $\sim$~1 with respect to
135: metallicity, and there is no signature of a primordial plateau abundance.
136:
137: These trends of Be and B abundances are explained by Galactic CR nucleosynthesis models of different types. One is the acceleration of
138: metal-rich CRs, probably freshly synthesized matter at SNe followed by the primary reactions between CR accelerated CNO nuclides and
139: interstellar nuclides $p$ and $\alpha$,
140: i.e. [CNO]$_{\rm CR}$+[$p\alpha$]$_{\rm ISM}
141: \rightarrow$[LiBeB]~\citep[e.g.][]{ram1997}. This mechanism leads to the same-rate increase of BeB and
142: O. \citet{fie2000} have suggested that the secondary reactions between CR accelerated $p\alpha$ and
143: interstellar CNO i.e. [$p\alpha$]$_{\rm CR}$+[CNO]$_{\rm ISM}
144: \rightarrow$[LiBeB], perhaps without any contribution of the primary
145: reactions, would explain the BeB to Fe trend if the ratio O/Fe increases
146: toward low metallicity. \citet{val2002} have shown that the
147: multi-zone (halo, thick disk, and thin disk) Galactic evolution model
148: including only the secondary reactions can reproduce the linear trend
149: without fine-tuning.
150:
151: In this work, I adopted Model 1 and the rapid burst model of \citet{dai2006} for
152: the cosmic chemical evolution model. This chemical evolution model
153: and other inputs, as well as calculation of light element evolution
154: are explained in Sec. 2. I present results of the CCR nucleosynthesis in
155: Sec. 3, and discuss this study in Sec. 4. I summarize the CCR
156: production of LiBeB in Sec. 5.
157:
158: \section{MODEL}
159:
160: \subsection{Cosmic SN Rate}
161:
162: I adopt the cosmic SFRs in a chemical evolution model
163: given by \citet{dai2006}. I take Model 1 and the rapid burst
164: model, which include formation of stars with masses between 40 and 100
165: $M_{\sun}$ in the early phase of the universe. As their best model, a parameter, the minimum mass $M_{\rm min}$ of dark matter halos of star-forming
166: structures is determined to be $10^7~M_{\sun}$ for chemical evolution of
167: structures. The adopted models are the same as~\citet{rol2006} use. The
168: birthrate function is given by
169: \begin{equation}
170: B(m,t,Z)=\phi_1(m)\psi_1(t)+\phi_2(m)\psi_2(Z),
171: \label{eq1}
172: \end{equation}
173: where $m, t, Z$ are the mass of star, the age of the universe, and the
174: metallicity, $\phi_1$ and $\phi_2$ are initial mass functions (IMFs) of
175: the normal and massive-only component of stars, respectively, $\psi_1$
176: and $\psi_2$ are SFRs for respective components.
177: The normal component is given such that its mass range is from 0.1 to
178: 100 $M_{\sun}$. The massive component is active only at high redshift,
179: and its mass range is from 40 to 100 $M_{\sun}$. The IMF of both modes
180: is given by a power law of mass with an index like the Salpeter's,
181: \begin{equation}
182: \phi_i(m)\propto m^{-(1+x)},
183: \label{eq2}
184: \end{equation}
185: with $x=1.3$. The amplitudes of the IMFs are normalized respectively as
186: \begin{equation}
187: \int_{m_{\rm inf}} ^{m_{\rm sup}} dm m \phi_i(m)=1,
188: \label{eq3}
189: \end{equation}
190: where $m_{\rm inf}$ and $m_{\rm sup}$ are the lower and upper ends for
191: the mass range of each mode.
192:
193: The normal-mode SFR is given by
194: \begin{equation}
195: \psi_1(t)=\nu_1 M_{\rm struct} \exp(-t/\tau_1),
196: \label{eq4}
197: \end{equation}
198: where $\nu_1=0.2$~Gyr~$^{-1}$ describes the efficiency of the star
199: formation, and $M_{\rm struct}$ and $\tau_1=2.8$~Gyr are mass of the
200: structure and timescale, respectively.
201:
202: In Model 1, the massive-mode SFR is given by
203: \begin{equation}
204: \psi_2(t)=\nu_2 M_{\rm ISM} \exp(-Z_{\rm IGM}/Z_{\rm crit}),
205: \label{eq5}
206: \end{equation}
207: where $\nu_2=80$~Gyr~$^{-1}$ is related to the efficiency of star
208: formation. $M_{\rm ISM}$ is the baryonic mass of the ISM. $Z_{\rm IGM}$ is the metallicity of the medium between
209: the collapsed structures (identified as the IGM), and
210: $Z_{\rm crit}=10^{-4} Z_\sun$~determines the effective epoch of the end
211: of Pop III star formation. In contrast, in the rapid burst
212: model,the massive mode star formation occurs as an instantaneous event
213: at redshift $z=16$. Since the SFR in the model has an instantaneous
214: bump and I found difficulty reading it from Fig. 13
215: of~\citet{dai2006}, I fixed the SFR so that it is $\sim 20 M_\sun {\rm
216: yr}^{-1} {\rm Mpc}^{-3}$ during $3\times 10^6$~yr~\citep{rol2006}.
217:
218: \subsection{LiBeB Production in the Homogeneous Universe}
219:
220: I calculate abundances of light elements (LiBeB) produced in the
221: homogeneous universe through the interaction between fast nuclei
222: accelerated in the SN shocks and background nuclei. The picture of
223: CCRs, i.e., the total kinetic energy of CRs and the propagation is identical to
224: that of~\citet{rol2006}.
225:
226: \subsubsection{CCR Energy and Its Spectrum}
227:
228: The total kinetic energy given to accelerated CRs by SN explosions is
229: \begin{eqnarray}
230: {\cal E}_{\rm SN}(z)&=& (1+z)^3 \nonumber \\
231: &\times&\int_{{\rm max}(8M_{\sun},m_{{\rm d}}(t))}^{m_{\rm sup}} dm \sum_{i=1}^2 \phi_i(m) \psi_i(t-\tau(m)) {\cal E}_{\rm CR}(m),\nonumber \\
232: \label{eq6}
233: \end{eqnarray}
234: where $m_{\rm d}(t)$ is the mass of stars with lifetime $t$, $\tau(m)$
235: is the lifetime of a star of mass $m$, and ${\cal E}_{\rm CR}(m)$ is the
236: energy imparted to CRs per SN with its initial mass
237: $m$. \citet{dai2004, dai2006} adopt stellar lifetimes $\tau(m)$ from
238: \citet{mae1989} for intermediate-mass stars ($<8~M_\sun$), and
239: those from \citet{sch2002} for massive stars ($8~M_\sun<m<100~M_\sun$).
240:
241: \citet{rol2006} give ${\cal E}_{\rm CR}(m)$ after some
242: assumptions to calculate the total kinetic energy ${\cal E}_{\rm
243: SN}(z)$. Core collapse SNe are supplied with energy by
244: the gravitational collapse of cores. Almost all of the energy generated
245: from core collapse, $E_{\rm CC}$ is transferred out by neutrinos. Only
246: 1~\% will be given to the energy of SN explosions.
247: \citet{rol2006} use a parameterization
248: \begin{equation}
249: {\cal E}_{\rm CR}(m)=\frac{\epsilon E_{\rm CC}(m)}{100},
250: \label{eq7}
251: \end{equation}
252: where $\epsilon$ is the fraction of SN explosion energy imparted
253: to CRs. Assumptions related to $E_{\rm CC}$ are as follows: Every star of mass $m>8~M_\sun$ explodes as SN. Stars of mass
254: $8~M_\sun < m < 30~M_\sun$ make neutron stars of mass 1.5~$M_\sun$ at
255: core collapses and $E_{\rm CC}=3\times 10^{53}$~ergs. Stars of mass
256: $30~M_\sun < m < 100~M_\sun$ become black holes with masses very similar
257: to their helium core mass~\citep{heg2003}. The mass of the helium
258: core is $M_{\rm He}=13(m-20~M_\sun)$~\citep{heg2002}. In this case
259: $E_{\rm CC}$ is proportional to the mass of the black hole, and
260: $E_{\rm CC}=0.3M_{\rm He}$ is assumed.
261:
262: I calculate light element production in the formalism of
263: \citet{mon1977} and \citet{rol2005, rol2006}. The proper source function of
264: SN CRs $Q_i(E,z)$ in a rapid burst at redshift $z_s$ (corresponding time
265: $t_s$) is defined as
266: \begin{eqnarray}
267: Q_i(E,z)=(1+z_s)^3 C(z_s)\frac{\phi_i(E,z_s)}{\beta}\delta(t-t_s)\nonumber \\
268: ~~~~~~~~~~~~~~~~~~~~~~~~[({\rm GeV/nucleon})^{-1}~{\rm cm}^{-3} {\rm s}^{-1}],
269: \label{eq8}
270: \end{eqnarray}
271: where $E$ and $\beta$ are the kinetic energy and velocity of CRs,
272: respectively, and I define the CR injection spectrum of nuclide $i$ as
273: \begin{equation}
274: \phi_i(E,z_s)=K_{ip}^{\rm CR}(z_s) \frac{1}{(E(E+2E_0))^{\gamma/2}},
275: \label{eq9}
276: \end{equation}
277: where $K_{ip}^{\rm CR}$ is the ratio of number abundance of $i$ to that
278: of $p$, i.e. $i/p$ of CRs, and $E_0=938$~GeV is the nuclear mass energy per
279: nucleon. The amplitude of the source function is set so that SNe from
280: both the normal mode (Pop II) and the massive mode (Pop III) stars supply the CR energy,
281: \begin{equation}
282: {\cal E}_{\rm SN}(z) = \int_{E_{\rm min}}^{E_{\rm max}} E\,\sum_i\,Q_i(E,z)\,dE.
283: \label{eq10}
284: \end{equation}
285: I take $E_{\rm min}=0.01$~MeV, $E_{\rm max}=10^6$~GeV, and CR spectral
286: index $\gamma=3$ as \citet{rol2006} do. Since the evolution of CR
287: confinement by a magnetic field is difficult to estimate~\citep{rol2006},
288: as a first step, I assume that the CR confinement is ineffective in the
289: early universe, so that all CRs generated by SNe in structures
290: immediately escape from structures to the IGM. In this case, there is
291: uniformity of the CR density in the universe.
292:
293: \subsubsection{Primary Light Element Production by SN CRs}
294:
295: I define the number density of a CR species $i$ of energy $E$ at
296: redshift $z$ as $N_i(E,z)$ [in cm$^{-3}$ (GeV/nucleon)$^{-1}$]. In
297: order to delete the volume changing effect by cosmic expansion, I define
298: the relative number abundance to that of the background proton $n_{\rm
299: H}(z)$,
300: \begin{equation}
301: N_{i,{\rm H}}(E,z)\equiv N_i(E,z)/n_{\rm H}(z).
302: \label{eq11}
303: \end{equation}
304: The transport equation for $N_{i,{\rm H}}$, under the isotropic condition, is~\citep{mon1977}
305: \begin{equation}
306: \frac{\partial N_{i,{\rm H}}}{\partial t} + \frac{\partial}
307: {\partial E}(bN_{i,{\rm H}}) + \frac{N_{i,{\rm H}}}{T_{\rm D}} = Q_{i,{\rm H}},
308: \label{eq12}
309: \end{equation}
310: where $b(E,z)\equiv (\partial E/\partial t)$ is the energy loss rate
311: [(GeV/nucleon)~s$^{-1}$] for cosmic expansion or ionization, and $T_{\rm
312: D}(E,z)$ is the
313: lifetime against destruction. $Q_{i,{\rm H}}(E,z)\equiv Q_i(E,z)/n_{\rm
314: H}(z)$ is the normalized comoving source function.
315:
316: The expansion loss and ionization loss are expressed in a product of
317: energy-dependent term and a redshift-dependent one, $b(E,z)=-B(E)f(z)$. These
318: terms are given in \citet{mon1977}. The redshift-dependent term of
319: the expansion loss is $f_{\rm E}=(1+z)^{-1} |dz/dt|H_0^{-1}$, where
320: $H_0$ is the Hubble constant. I assume the standard $\Lambda$CDM model
321: with its parameters from WMAP three-year data~\citep{spe2007}\footnote{http://lambda.gsfc.nasa.gov} ,
322: $h=H_0/(100~{\rm km~s}^{-1}~{\rm Mpc}^{-1})=0.704$, $\Omega_b
323: h^2=0.022$, $\Omega_m =0.27$, $\Omega_\Lambda =0.73$. The ionization
324: loss rate is from the fitting formula in \citet{men1971} with
325: the number fraction of $^4$He, He/H=0.08. $T_{\rm D}=(n_{\rm H}(z)
326: \sigma_{{\rm D},i} \beta)^{-1}$ is estimated with the nuclear destruction
327: cross section $\sigma_{{\rm D},i}$ from \citet{ree1974}.
328:
329: I define $z^\star(E,E',z)$ as in \citet{mon1977},
330: \begin{equation}
331: \frac{\partial z^\star}{\partial E}=-\frac{1}{B(E)f(z)}\left|\frac{dz}{dt}\right|\frac {\partial z^\star}{\partial z}.
332: \label{eq13}
333: \end{equation}
334: A physical interpretation is that a CR particle with energy $E$ at
335: redshift $z$ had an energy $E'(\geq E)$ at redshift $z^\star(E,E',z)$
336: before experiencing energy loss. Thus $z^\star(E,E,z)=z$ is satisfied.
337: CCR particles with energy $E$ at $z$ originate in those with
338: $E'_s(E,z,z_s)$ at $z_s$. $E'_s(E,z,z_s)$ satisfies an equation,
339: $z^\star(E,E'_s,z)=z_s$. $z^\star(E,E',z)$ is obtained by integrating
340: Eq.~(\ref{eq13}) applying the greater loss process to $b$
341: assuming that the process with the greater rate of $b(E,z)$ is dominant all
342: the way from redshift $z^\star$ to $z$.
343:
344: The transfer equation is solved~\citep{mon1977} to obtain the CCR
345: energy spectrum from a CR burst at $z_s$,
346: \begin{eqnarray}
347: \hspace{-10pt}\Phi_{i, {\rm H}}(E,z,z_s) &=& C(z_s) \frac{\phi_i(E'_s,z_s)}{n_{\rm H}^0}\frac{\beta}{\beta'} \left|\frac{d{z}}{d{t}}\right|_{z_s}\nonumber \\
348: &&\times\frac{\exp{(-\xi(E,E'_s,z))}}{|b(E,z_s)|}\,\frac{1}{\left|\partial z^\star/\partial E'\right|_{{E'_s}}},
349: \label{eq14}
350: \end{eqnarray}
351: where $\Phi_{i,{\rm H}}(E,z,z_s)\equiv \Phi_i(E,z,z_s)/n_{\rm H}(z)$ is
352: the normalized flux of $i$ per comoving volume with $\Phi_i(E,z,z_s)\equiv
353: \beta N_i(E,z)_{z_s}$, $\beta$ and $\beta'$ are the velocities corresponding
354: to energy $E$ and $E'_s$, respectively. $n_{\rm H}^0$ is the present
355: average number density of protons in the universe. $\xi$ is an effect
356: resulting when the nuclear destruction is considered, and given as
357: \begin{equation}
358: \xi(E,E'_s,z)\hspace{-2pt}
359: =\hspace{-4pt}
360: \int_E^{E'_s}\hspace{-4pt}
361: \frac{dE''}{\left|b(E'',z^\star(E,E'',z)) T_{\rm D}(E'',z^\star(E,E'',z))\right|}.
362: \label{eq15}
363: \end{equation}
364: After analysis with Eq.~(\ref{eq13}), one can estimate $|\partial
365: z^\star / \partial E'|_{E'=E'_s}=|b(E'_s,z_s)|^{-1} |dz/dt|_{z=z_s}$ and
366: find an expression for $\Phi_{i, {\rm H}}$
367: \begin{equation}
368: \Phi_{i, {\rm H}}(E,z,z_s) = C(z_s) \frac{\phi_i(E'_s,z_s)}{n_{\rm H}^0}\frac{\beta}{\beta'} \frac{|b(E'_s,z_s)|}{|b(E,z_s)|} e^{-\xi(E,E'_s,z)}.
369: \label{eq16}
370: \end{equation}
371:
372: The production rate of light element $l$ of energy $E$, produced at
373: redshift $z$ is given by
374: \begin{eqnarray}
375: \hspace{-4pt}\frac{\partial N_{l,\,{\rm H}}(E,z,z_s)}{\partial t}\hspace{-2pt} &=& \sum_{i,j} \hspace{-3pt}
376: \int \hspace{-4pt}
377: \sigma_{ij \rightarrow l}(E,E')n_j(z)\Phi_{i,{\rm H}}(E',z,z_s)\, dE' \nonumber \\
378: &=& \sum_{i,j} \hspace{-3pt}
379: \int \hspace{-4pt}
380: \sigma_{ij \rightarrow l}(E,E') K_{jp}^{\rm IGM}(z)\Phi_i(E',z,z_s)\, dE',\nonumber \\
381: \label{eq17}
382: \end{eqnarray}
383: where $\sigma_{ij \rightarrow l}(E,E')$ is a cross section of a process between
384: a CR nuclide $i$ with energy per nucleon $E'$ and a background species $j$ to make a given light
385: element $l$ with $E$, and $n_j(z)$ and $K_{jp}^{\rm IGM}(z)$ are
386: background number abundance of a nuclide $j$ and number ratio of $j$ to
387: proton, respectively. When the destruction of the light element $l$
388: after production is neglected, the total production rate is calculated as
389: \begin{eqnarray}
390: &&\int \frac{\partial N_{l,\, {\rm H}}(E,z,z_s)}{\partial t} dE \nonumber\\
391: &&= \sum_{i,j} K_{jp}^{\rm IGM}(z) \int \sigma_{ij \rightarrow l}^{\rm tot}(E') \Phi_i(E',z,z_s)\, dE',
392: \label{eq18}
393: \end{eqnarray}
394: where $\sigma_{ij \rightarrow l}^{\rm tot}(E')$ is the total cross section
395: of a reaction $i+j \rightarrow l+X$, with any $X$. I adopt cross sections
396: from \citet{rea1984}, and particularly for the $\alpha+\alpha$ reaction,
397: exponential-plus-constant cross section for $l=^6$Li and exponential one
398: for $l=^7$Li from \citet{mer2001}. The resulting light element
399: abundance is obtained as the CR production added to the BBN yield. The
400: yield by CR nucleosynthesis is the integration of those produced at $z'$
401: from CRs generated at $z_s$ over $z'$ and $z_s$, thus
402: \begin{eqnarray}
403: \left(\frac{l}{\rm H}\right)_{\rm IGM}\hspace{-15pt}(z)&=& \left(\frac{l}{\rm H}\right)_{\rm BBN} \nonumber \\
404: &&+\int_z^{z_{\rm max}} dz_s \left|\frac{dt}{dz_s}\right| \int _z ^{z_s} dz' \left|\frac{dt}{dz'}\right| \nonumber \\
405: &&\times \sum_{i,j} K_{jp}^{\rm IGM}(z') \int \sigma_{ij \rightarrow l}^{\rm tot}(E') \Phi_i(E',z',z_s)\, dE'.\nonumber \\
406: \label{eq19}
407: \end{eqnarray}
408:
409: \subsubsection{Secondary Light Element Production by SN CRs}
410:
411: I also calculate the LiBeB production in the universe by the secondary
412: process, i.e., [$p\alpha$]$_{\rm CR}$+[CO]$_{\rm
413: ISM}\rightarrow$[LiBeB]$_{\rm ISM}$. Since the C and O abundances of
414: the ISM in structures are about two orders of magnitude higher than those
415: of the IGM~\citep[see Fig. 11 in][]{dai2006}, the secondary LiBeB
416: production in the IGM is not important. I expect that the LiBeB
417: abundances in the ISM are enhanced by a contribution of the secondary
418: process. In fact, the reactions of [$p\alpha$]$_{\rm CR}$+[CO]$_{\rm
419: ISM}\rightarrow$[LiBeB]$_{\rm ISM}$ make light elements in the ISM and
420: the mass accretion to the structures from the IGM dilutes the
421: ISM abundances in the framework of this model involving a hierarchical
422: structure formation. Note that from the assumption that the confinement of CRs by
423: a magnetic field is ineffective, the CRs do not stay in the structures.
424:
425: The light element abundances
426: produced by the secondary reactions are then given with a parameter: the
427: fraction of baryons at redshift $z$ which are in structures
428: \begin{equation}
429: f(z) = \frac{\int_{M_{\rm min}}^{\infty}~dM~M~f_{\rm PS}(M,z)}{\rho_{\rm DM}},
430: \label{eq20}
431: \end{equation}
432: where $f_{\rm PS}(M,z)$ is the distribution function of
433: halos taken
434: from the \citet{she1999} modification to the Press-Schechter
435: function~\citep{pre1974} converted into the mass function~\citep{jen2001} by a code provided by
436: A. Jenkins (2007, private communication). I assume that the primordial
437: power spectral slope is $n$=1, the rms amplitude for mass density
438: fluctuations in a sphere of radius 8 $h^{-1}$ Mpc is $\sigma_8=0.9$, and
439: the \citet{bon1984} fit to the transfer function for cold dark
440: matter is used in generating a mass function. $\rho_{\rm DM}$ is the
441: comoving dark matter density of the universe.
442:
443: The light elements made by the secondary process are contained in the
444: structures that grow gradually. The abundance of a light element in the
445: structures is then given by
446: \begin{eqnarray}
447: \left(\frac{l}{\rm H}\right)_{\rm ISM}\hspace{-15pt}(z)&=& \left(\frac{l}{\rm H}\right)_{\rm IGM}\hspace{-15pt}(z) \nonumber \\
448: &&\hspace{-5pt}+ \frac{1}{f(z)}\int_z^{z_{\rm max}} dz_s \left|\frac{dt}{dz_s}\right| \int _z ^{z_s} dz' \left|\frac{dt}{dz'}\right| \nonumber \\
449: &&\hspace{-5pt}\times \sum_{i,j} K_{jp}^{\rm ISM}(z') f(z') \int \sigma_{ij \rightarrow l}^{\rm tot}(E') \Phi_i(E',z',z_s)\, dE'.\nonumber \\
450: \label{eq21}
451: \end{eqnarray}
452:
453: \section{RESULTS}
454:
455: I calculate the light element production in the uniform universe by CCRs
456: (i.e. neglecting an inhomogeneity of CCRs). I consider only processes
457: between the accelerated CRs with the abundance patterns of the structures
458: in the~\citet{dai2006} model and the background IGM abundance which I assumed
459: to be of the primordial abundance. I set the primordial helium abundance He/H=0.08.
460:
461: I show the result of light element evolution in Model 1 in Fig.\
462: \ref{fig1}. Solid lines correspond to the case where both the normal
463: and massive modes are included, and dashed lines correspond to the case
464: in which only the normal mode is included as the energy source. When the SN
465: energy is totally given to the CR acceleration ($\epsilon=1$),
466: $^6$Li/H=$2.0\times 10^{-11}$ at $z=3$ is derived. In
467: Fig.\ \ref{fig1}, $\epsilon=0.31$ is assumed so that $^6$Li/H=$6\times
468: 10^{-12}$ at $z=3$ results that is the observed abundance level in
469: MPHSs. In the case that the CRs are energized by only the normal mode star
470: formation, the energy fraction $\epsilon=0.73$ is needed to realize the
471: observed $^6$Li abundance. The time evolution of light element
472: abundances in the rapid burst model is shown in Fig.\ \ref{fig2}. The
473: energy fraction given to the CR acceleration is assumed to be
474: $\epsilon=0.029$, and the abundance of $^6$Li gets $^6$Li/H=$6\times
475: 10^{-12}$ at $z=3$. These results are very similar to that
476: in~\citet{rol2006} (See their Fig. 2). In Model 1, $^6$Li is produced
477: gradually with decreasing redshift, while in rapid burst model, it is
478: immediately produced by most part at the burst of star formation. In my
479: calculation, the $^6$Li production at high redshift tends to be slightly
480: more efficient than in~\citet{rol2006}. This difference might be caused
481: by different numerical calculation of transport equation for nuclides.
482:
483: \placefigure{fig1}
484: \placefigure{fig2}
485:
486: Figure\ \ref{fig3} shows the yields by CCRs generated at $z_s$ per second
487: in Model 1, i.e. $\Delta$($l$/H)/$\Delta t_s$. The calculated
488: results show that $^6$Li and $^7$Li are produced mainly by the $\alpha+\alpha$ fusion
489: reaction, and Be and B production has the strongest contribution from
490: the O+$p$ (and C+$p$) spallation processes. I have taken the helium
491: abundance He/H=0.08, and the O abundance in the ISM of structures in
492: Model 1 is roughly constant from $z\sim 0$ to $z\sim 20$~\citep[see Fig. 11 of][]{dai2006}. The injected energy density in CRs is smoothly
493: increasing as a function of redshift $z$~\citep[Fig. 2 of][]{rol2006}. This figure therefore reflects the abundance of seed
494: nuclides and injected energy density. The CRs generated at $z_s\sim 3$
495: have little time to react with background nuclear species, so that the
496: yields go to zero, and a decrease in all yields at $z_s\lesssim 25$
497: reflects the shape of CR energy density~\citep[Fig. 2 of][]{rol2006}.
498:
499: \placefigure{fig3}
500:
501: I assume that MPHSs formed at redshift $z\sim 3$ and that they include
502: LiBeB elements at the level of the IGM abundances at the time. If one can
503: neglect the inhomogeneity of CCRs resulting from the local growth of a
504: magnetic field, numbers of produced elements are proportional to target
505: particle numbers. Consequently, resulting light element abundance $l$/H
506: does not depend on density there. Although the metallicities like Fe/H
507: of metal-poor stars reflect how the metal-enhanced and metal-deficient
508: gas are mixed before star formation, the light element abundances are
509: constant for the same formation epoch. In this case, there appear
510: primordial plateaus on the plot of abundances as a function of
511: metallicity [Fe/H]. The energy fractions given to the CR acceleration,
512: $\epsilon=0.31$ for Model 1 and $\epsilon=0.029$ for the rapid burst
513: model realize the observed abundance of $^6$Li in MPHSs at $z=3$. These
514: fractions are reasonable, considering that the energy of Galactic CRs can
515: be covered by $10-30$ percent of supernova remnant (SNR) energy~\citep{dru1989}. If
516: $^6$Li observed in MPHSs is produced mainly by this CCR nucleosynthesis,
517: these stars include Be and B which had been coproduced.
518:
519: In Fig.\ \ref{fig4} the plateau levels of $^6$Li and $^7$Li in Model 1
520: are drawn with observational data points in a plot of abundances as a
521: function of metallicity [Fe/H]. As for the $^7$Li abundance, the BBN
522: prediction is calculated using the Kawano code~\citep{kawano1992} with the
523: use of the new world average of the neutron lifetime~\citep{mat2005}. I take the energy density of baryons in the universe given
524: by WMAP first year data analysis~\citep{spe2003} that is $\Omega_b
525: h^2=0.0224\pm 0.0009$. This value corresponds to the
526: baryon-to-photon ratio $\eta=6.1\times 10^{-10}$. The predicted
527: abundance is then ($^7$Li/H)$_{\rm BBN}$=4.5$\times 10^{-10}$. The
528: $^7$Li production by the CCR nucleosynthesis is a minor addition to the
529: BBN result. The figure for the rapid burst model is very similar to
530: Fig.\ \ref{fig4}. The factor of about three difference between the BBN
531: prediction and the observation of $^7$Li abundance is apparent.
532:
533: \placefigure{fig4}
534:
535: The coproduced abundance levels of Be and B in Model 1 are shown in Fig.\ \ref{fig5}
536: with observational data points. It is interesting that the predicted
537: abundance levels of Be and B are located at nearly the lowest point ever
538: detected. This model predicts primordial plateau abundances of Be and B
539: as an analog of the likely $^6$Li plateau. If future observation of
540: beryllium in metal-poor stars catches evidence of a plateau abundance,
541: it would be explained by this CCRs origin. When the normal mode star
542: formation alone is taken as the CR energy source, the result is consistent
543: with the observed abundance of $^6$Li in MPHSs at $z=3$ with
544: $\epsilon=0.73$. This case corresponds to dashed lines in Fig.\
545: \ref{fig5}. In the same way, the coproduced abundances in the rapid
546: burst model with $\epsilon=0.029$ are shown as the dotted lines. This calculation shows that Be and B abundances are somewhat
547: higher in the rapid burst model than in Model 1 relative to that of
548: $^6$Li. This calculation seems possibly to contain an error of the
549: order of $\sim 10$~\% in Be and B abundances associated with the
550: reading of a very brief burst of SFR~\citep[Fig. 13 of][]{dai2006} and
551: resulting sudden enrichment of C and O in the ISM of
552: structures~\citep[Fig. 16 of][arXiv:astro-ph/0509183]{dai2006} as input
553: profiles in this calculation. In the cosmic chemical evolution model
554: of~\citet{dai2006}, the rapid burst of star formation occurs in a very
555: brief time and C and O abundances in structures increase accordingly
556: reflecting the ejection of C and O nuclei by SNe. Since Be and B
557: production is sensitive to the CR energy density times C and O
558: abundances in structures, the fine-meshed time evolution of C and O
559: abundances as well as CR energy density (which is associated with SN
560: rate) is necessary to perform a precise calculation. However, since the rapid burst model gives sudden rises of C and O
561: abundances, and C and O particles are accelerated before the ISM
562: abundances can be somewhat diluted by accretion of the IGM, abundances of C and O
563: in CCRs are relatively high. Then it is as expected that the rapid
564: burst model produces more Be and B relative to $^6$Li than in Model 1,
565: considering that Be and B are produced by spallation processes of C and
566: O while lithium isotopes are produced mainly by $\alpha+\alpha$ fusion.
567: I show the results of calculations of light element production in
568: the CCR nucleosynthesis model in Table \ref{tab1}. In the second
569: column, values of energy fraction of SNRs given to CR acceleration
570: required to produce $^6$Li at the observed level in MPHSs at $z=3$ are
571: shown. The resulting abundances of light elements with the CR
572: acceleration efficiencies in the second column are listed in the third
573: to seventh columns.
574:
575: \placefigure{fig5}
576: \placefigure{fig6}
577: \placetable{tab1}
578:
579: In the calculations so far, the spectral index of the CR injection
580: spectrum has been fixed to $\gamma=3$. I show the energy fraction of
581: SNRs to CRs, $\epsilon$, required to produce $^6$Li at the MPHSs level at
582: $z=3$ as a function of $\gamma$ in Fig.\ \ref{fig6}. The solid line
583: corresponds to Model 1 and the dashed line to the case where only the
584: normal mode stars are considered as the energy source of CCRs. One can
585: see that the required energy fraction $\epsilon$ is reasonable if
586: $2.7\lesssim \gamma \lesssim 3.1$. Meanwhile, smaller spectral index of
587: $\gamma \lesssim 2.7$ could not produce enough $^6$Li in this model.
588:
589: \placefigure{fig7}
590:
591: Figure\ \ref{fig7} shows abundances of Be and B at $z=3 $ as a
592: function of $\gamma$ in Model 1 with the value of $\epsilon$
593: in Fig.\ \ref{fig6}, i.e., when $^6$Li is produced at the MPHSs level. It is
594: found that relative produced abundances of Be and B are decreasing
595: functions of $\gamma$. In other words, the steeper the momentum spectrum
596: of CRs is, the more $^6$Li is produced relatively. In the region
597: $2.7\lesssim \gamma \lesssim 3.1$, where the $^6$Li abundance produced
598: in this model can attain the MPHSs level with a reasonable partition
599: of SNR energy to CR acceleration, log(Be/H)$\sim -13.1$~-~$-12.9$ and
600: log(B/H)$\sim -12.0$~-~$-11.8$ are obtained. These abundances are again
601: near the least abundances detected in metal-poor stars.
602:
603: \placefigure{fig8}
604:
605: Figure\ \ref{fig8} shows the abundances of light elements produced by the
606: secondary process in the ISM as a function of redshift in Model 1
607: calculated by the second term in the rhs of Eq.\
608: (\ref{eq21}). $\epsilon=0.31$ is assumed to result in $^6$Li/H=$6\times
609: 10^{-12}$ in the IGM at $z=3$. Solid lines correspond to the case where
610: both the normal and massive modes are included, and dashed lines
611: correspond to a contribution of the normal mode as the energy source.
612: The contributions of primary and secondary processes to the light element
613: synthesis in the ISM are estimated by the comparison between Fig.\
614: \ref{fig1} and Fig.\ \ref{fig8}. Since $^6$Li and $^7$Li are produced mainly
615: by the $\alpha+\alpha$ fusion which I contained in the calculation of
616: Fig.\ \ref{fig1}, the contributions of the secondary process are relatively
617: small. On the other hand, Be and B have some contribution from the
618: secondary process. For example, the Be and B abundances produced by the
619: secondary process are about 25~\% (65~\%) of those by the primary
620: process at $z=3$ in Model 1 with both the normal and massive modes
621: (with only the normal mode). It is found that the BeB abundance levels
622: produced by the CCR nucleosynthesis are roughly the same in the ISM and
623: the IGM. This contribution from the secondary
624: process is small in the rapid burst model (about 2.5~\% for Be and B) at
625: $z=3$. The fraction of baryons included in the ISM is smaller
626: at higher redshift, when the energy injection in the rapid burst model
627: occurs. It is then easily understood that the light element abundances
628: produced by the secondary reactions are smaller in the rapid burst model
629: considering the dilution of light element abundances due to mass
630: accretion by structure formation in the model.
631:
632: \section{DISCUSSION}
633: The resulting abundances from Eq.\ (\ref{eq18}) have a clear dependence on
634: input quantities as
635: \begin{equation}
636: \int \frac{\partial N_{l,\, {\rm H}}(E,z,z_s)}{\partial t} dE \propto \sum_{i,j} {\cal E}_{\rm SN}(z_s) K_{ip}^{\rm CR}(z_s) K_{jp}^{\rm IGM}(z).
637: \label{eq22}
638: \end{equation}
639: I adopted the results of the two models of \citet{dai2006} as the total
640: CRs energy by SN explosions ${\cal E}_{\rm SN}(z_s)$ and the ratio of
641: number abundance of $i$ to that of $p$, i.e. $i/p$ of CRs $K_{ip}^{\rm CR}(z_s)$,
642: which I assume to be the same as the ISM abundance of structures.
643: These two quantities are uncertain, in fact. I use the primordial
644: abundance for the background number ratio of $j$ to $p$ $K_{jp}^{\rm
645: IGM}(z)$. This quantity is reasonable and would not contain large
646: uncertainty. A large difference in $K_{ip}^{\rm CR}(z_s)$ ($i$=C,O) leads to a
647: difference in produced Be and B abundances linearly, while $^6$Li and
648: $^7$Li abundances have little influence since they are mainly produced by
649: the $\alpha+\alpha$ fusion process.
650:
651: \citet{suz2001} have developed a model for the evolution of
652: light elements in the Galaxy, which includes the SN-induced
653: chemical evolution with contributions from SNe and CRs nucleosynthesis
654: self-consistently. They have explained the light element abundances
655: observed in metal-poor stars using their model with the CR abundances
656: including SN ejecta 3.5~\% in mass fraction. The linear relation
657: between [BeB/H] and [Fe/H] has been obtained by the primary process to
658: make Be and B. Consequently their CR abundance includes the C and O in
659: mass fraction of $(5-7)\times 10^{-3}$ originating from the SN ejecta
660: throughout the Galactic chemical evolution. On the other hand, the mass
661: fraction of C+O is $(3-10)\times 10^{-3}$ in Model 1 of \citet{dai2006} I adopt, which is roughly the same level of abundance as
662: that \citet{suz2001} used within a factor of $\lesssim 2$. The
663: CR abundance I give in this study, therefore, would be appropriate
664: within a factor of $\sim 2$, even if the real C and O abundances in the
665: ISM producing CRs were lower than I give, supposing that the CRs have
666: contribution from the SN ejecta by the fraction inferred
667: by~\citet{suz2001}.
668:
669: I assume that all CRs escape from structures to the IGM, and do not
670: consider a nonuniformity of the CR density in the universe. As
671: structures grow in the universe, a magnetic field grows accordingly, and
672: the CR flux might get inhomogeneous, especially at low redshift, while the overproduction of $^6$Li provides a constraint on the
673: confinement of CRs in the ISM~\citep{rol2006}. Further study
674: including the space distribution and time evolution of a magnetic field
675: is desirable to estimate the light element abundances produced by the CCR
676: nucleosynthesis.
677:
678: The calculated result would also contain an uncertainty from two-step
679: reactions, i.e., production of nuclide $l_2$ by sequential non-thermal
680: nuclear reactions: $i+j\rightarrow l_1$ and $l_1+j_2\rightarrow l_2$ with
681: any nuclides $l_1$ and $j_2$~\citep{ram1997,kne2003}. \citet{kne2003}
682: have found that two-step reaction rates are only of the order of 1/10
683: smaller than one-step reaction rates. The effect of two-step reactions
684: in the framework of this study is roughly estimated as follows. Nuclei produced by nuclear reactions
685: experience an energy loss and nuclear destruction. If one take nuclei
686: with kinetic energy of $E\gtrsim 10$~MeV/nucleon, the expansion loss is
687: dominant loss process in the considered redshift range. The time scale
688: for expansion loss is given by
689: \begin{eqnarray}
690: T_{\rm loss}&\sim& \frac{dE}{b_{\rm exp}(E,z)}\nonumber \\
691: &=&\frac{E+E_0}{H_0 (E+2E_0)}\left[\Omega_m(1+z)^3 + (1-\Omega_m)\right]^{-1/2} \nonumber \\
692: &=&4.4\times 10^{17}~{\rm s} \frac{E+E_0}{(E+2E_0)}\left[\Omega_m(1+z)^3 + (1-\Omega_m)\right]^{-1/2},\nonumber \\
693: \label{eq23}
694: \end{eqnarray}
695: where $b_{\rm exp}(E,z)$ is the energy loss rate for the cosmic
696: expansion. On the other hand, using an empirical formula of nuclear
697: destruction~\citep[Eq.\ 1 in ][]{let1983}, i.e., $\sigma=44.9
698: A^{0.7}$~mb, the time scale for nuclear destruction is given by
699: \begin{eqnarray}
700: T_{\rm D}&=& (n_{\rm H}(z) \sigma \beta)^{-1} \nonumber \\
701: &=&8.0\times 10^{20}~{\rm s} \left(\frac{A}{10}\right)^{-0.7} (1+z)^{-3} \beta^{-1}.
702: \label{eq24}
703: \end{eqnarray}
704: Ratios of time scales for nuclides with $A\sim 10$ and
705: $E=10-10^6$~MeV/nucleon are $T_{\rm loss}/T_{\rm D}\sim 10^{-3}-10^{-2}$
706: at $z=3$, $3\times 10^{-3} - 3\times 10^{-2}$ at $z=10$, and $10^{-2} -
707: 0.2$ at $z=30$, respectively. Therefore, the fraction at which nuclei
708: produced by nuclear spallations experience second nuclear reactions
709: before losing enough energy is expected to be at most of the order of
710: $\sim 10$~\%, if the energy spectrum of CRs indicates small amount of
711: high energy CRs. I check a fraction of high energy CRs in number, which
712: would be estimated by
713: \begin{equation}
714: F(\gamma)\equiv \frac{\int_{\rm 100~MeV}^{E_{\rm max}} Q_i(E',z)dE'}{\int_{\rm 10~MeV}^{E_{\rm max}} Q_i(E',z)dE'}=\frac{P({\rm 100~MeV})}{P({\rm 10~MeV})},
715: \label{eq25}
716: \end{equation}
717: where I defined
718: \begin{equation}
719: P(E)\equiv \int_E^{E_{\rm max}} \frac{E'+E_0}{\left[E'(E'+2E_0)\right]^{(\gamma+1)/2}} dE'.
720: \label{eq26}
721: \end{equation}
722: For $\gamma=2, 3, 4$, $F(2)=0.31$, $F(3)=0.095$, and $F(4)=0.029$ are
723: obtained, respectively. The fraction of high energy CRs is thus
724: relatively low for the range of $2\leq \gamma \leq 4$. As a
725: result, the effect of two-step reactions is small and would be at most of the order
726: of $\sim 10$~\%.
727:
728: \citet{rol2008} also study the CR production of Be and B by CCRs. Their
729: conclusion is very similar to that of this study, and a potentially
730: detectable Be and B is produced by CCR-induced spallation reactions at
731: the time of the formation of the Galaxy ($z\sim 3$). However, there are
732: some differences of assumptions between the two studies, which are
733: compared here. I give the CR spectrum of CO nuclides by the same shape
734: as those of $p$ and $\alpha$ particles, while~\citet{rol2008} give it by
735: a broken power law which matches the observed present-day Galactic CR
736: spectrum. Moreover, I give the abundances of CR by those of structures
737: in the cosmic chemical evolution model of~\citet{dai2006},
738: while~\citet{rol2008} give them by those of structures
739: in~\citet{dai2006} model multiplied by abundance enhancement factors of
740: present Galactic CR fluxes. I assume that the CR confinement by a
741: magnetic field is ineffective in the early universe, and that all CRs
742: generated by SNe in structures escape to the IGM, while~\citet{rol2008}
743: apply a shape of CR diffusion coefficient, which leads to some degrees
744: of the nuclear destruction of CRs in structures and the application of a
745: broken power law in the CR energy spectrum of CO nuclides. Their
746: diffusion coefficient is derived assuming that it is given by that in a
747: magneto-hydrodynamic (MHD) turbulence, and that the magnetic energy
748: density is proportional to the thermal energy, and that a characteristic
749: length scale for the magnetic field is given by the local Jeans scale.
750: It is interesting that the two studies using different assumptions for
751: the uncertain physical inputs arrived at the similar conclusions.
752:
753: I comment on a difference between the results of this CCR production of
754: the light elements and those of the flare production model~\citep{tat2007}. The CCR nucleosynthesis leads to the production of Be
755: and B at the lowest level ever detected, in this model calculation. The
756: nucleosynthesis on the main sequence stars triggered by flare-accelerated nuclides, on the other
757: hand, results in the negligible productions of $^7$Li, Be and B, which
758: are proportional to the metallicity and exist at very low abundance
759: levels under the observed linear trend as a function of metallicity.
760: Therefore if we observe a signature of the primordial origin of Be and B
761: by measurements of MPHSs, the production mechanism of Be and B would not
762: be the flare-energized nuclear reactions after the star formations, and
763: it would be thought that the plateau abundances of Be and B originate
764: in the CCR production and $^6$Li has been coproduced at pregalactic
765: phase by the CCRs.
766:
767: \section{CONCLUSIONS}
768: The recent observations of MPHSs reveal the probable existence of high
769: plateau abundance of $^6$Li, which is about a thousand times higher than
770: predicted in the standard BBN model. Since the standard Galactic
771: chemical evolution model with the Galactic CR nucleosynthesis
772: gives lower values of $^6$Li abundance at the metallicities of the observed MPHSs, some mechanism must have produced $^6$Li existing
773: in the surface of MPHSs. As a candidate of a $^6$Li production
774: mechanism, the early burst of CRs has been proposed~\citep{rol2005}, and the nucleosynthesis by the CRs from SN
775: explosions is calculated~\citep{rol2006} in a detailed model of
776: cosmic chemical evolution~\citep{dai2006} which satisfies various observational
777: constraints including an early reionization of the universe. \citet{rol2006} have found that the $\alpha+\alpha$
778: fusion reaction can produce $^6$Li to the level observed in MPHSs.
779:
780: I calculate the cosmological cosmic ray nucleosynthesis of Be and B
781: isotopes as well as $^6$Li and $^7$Li with the use of Model 1 and
782: the rapid burst model in
783: \citet{dai2006}. It is assumed that all CRs produced by SNe in the
784: ISM escape to the IGM and the CR intensity is always homogeneous in the
785: universe. I found that when Model 1 (the rapid burst model) of \citet{dai2006} is adopted for the SFR of the universe and the metal
786: abundances of CRs, Be and B are produced at (above) the levels observed
787: in the most metal-poor stars with detection of Be or B, if the $^6$Li
788: plateau abundance is made by the same CCR nucleosynthesis. The CR
789: acceleration energy needed to make $^6$Li primordial plateau abundance
790: at the observed level is $\sim 3-31$~\% of the SN kinetic energy. This
791: value is not too large in view of an inferred present fraction of the SN energy
792: used to the CR acceleration~\citep{dru1989}. The pregalactic SN
793: activity might have produced some level of light elements. Although the
794: resulting abundances of light elements depend on the parameters which I
795: fixed to the values of \citet{dai2006}, the future measurements of
796: metal-poor stars would show the reasonableness of this early $^6$LiBeB
797: production mechanism, and might provide a signature of a primordial Be
798: (and perhaps B) plateau. Further observations of LiBeB elements in
799: MPHSs are highly desirable and valuable.
800:
801: \acknowledgments
802:
803: I thank Fr\'ed\'eric~Daigne for explaining the result of his chemical
804: evolution calculation, and Emmanuel~Rollinde for helpful comments. I
805: am grateful to Adrian~Jenkins for giving me his code to make the mass
806: function of dark matter halos. I appreciate the support of the Japan
807: Society for the Promotion of Science.
808:
809: \begin{thebibliography}{}
810: \bibitem[Asplund et al.(2006)]{asp2006} Asplund, M., Lambert,
811: D.~L., Nissen, P.~E., Primas, F., \& Smith, V.~V.\ 2006, \apj, 644, 229
812:
813: \bibitem[Boesgaard et al.(1999)]{boe1999} Boesgaard, A.~M.,
814: Deliyannis, C.~P., King, J.~R., Ryan, S.~G., Vogt, S.~S., \& Beers, T.~C.\
815: 1999, \aj, 117, 1549
816:
817: \bibitem[Boesgaard \& Novicki(2006)]{boe2006} Boesgaard, A.~M.,
818: \& Novicki, M.~C.\ 2006, \apj, 641, 1122
819:
820: \bibitem[Bond \& Efstathiou(1984)]{bon1984} Bond, J.~R., \&
821: Efstathiou, G.\ 1984, \apjl, 285, L45
822:
823: \bibitem[Bonifacio et al.(2007)]{bon2007} Bonifacio, P., et al.\ 2007,
824: \aap, 462, 851
825:
826: \bibitem[Coc et al.(2004)]{coc2004} Coc, A., Vangioni-Flam, E.,
827: Descouvemont, P., Adahchour, A., \& Angulo, C.\ 2004, \apj, 600, 544
828:
829: \bibitem[Cumberbatch et al.(2007)]{cum2007} Cumberbatch, D.,
830: Ichikawa, K., Kawasaki, M., Kohri, K., Silk, J.,
831: \& Starkman, G.~D.\ 2007, \prd, 76, 123005
832:
833: \bibitem[Cunha et al.(2000)]{cun2000} Cunha, K., Smith, V.~V.,
834: Boesgaard, A.~M., \& Lambert, D.~L.\ 2000, \apj, 530, 939
835:
836: \bibitem[Daigne et al.(2006)]{dai2006} Daigne, F., Olive,
837: K.~A., Silk, J., Stoehr, F., \& Vangioni, E.\ 2006, \apj, 647, 773
838:
839: \bibitem[Daigne et al.(2004)]{dai2004} Daigne, F., Olive,
840: K.~A., Vangioni-Flam, E., Silk, J., \& Audouze, J.\ 2004, \apj, 617, 693
841:
842: \bibitem[Drury et al.(1989)]{dru1989} Drury, L.~O.,
843: Markiewicz, W.~J., \& Voelk, H.~J.\ 1989,
844: \aap, 225, 179
845:
846: \bibitem[Duncan et al.(1997)]{dun1997} Duncan, D.~K., Primas,
847: F., Rebull, L.~M., Boesgaard, A.~M., Deliyannis, C.~P., Hobbs, L.~M., King,
848: J.~R., \& Ryan, S.~G.\ 1997, \apj, 488, 338
849:
850: \bibitem[Fields et al.(2000)]{fie2000} Fields, B.~D., Olive,
851: K.~A., Vangioni-Flam, E., \& Cass{\'e}, M.\ 2000, \apj, 540, 930
852:
853: \bibitem[Garcia Lopez et al.(1998)]{gar1998} Garcia Lopez,
854: R.~J., Lambert, D.~L., Edvardsson, B., Gustafsson, B., Kiselman, D., \&
855: Rebolo, R.\ 1998, \apj, 500, 241
856:
857: \bibitem[Heger et al.(2003)]{heg2003} Heger, A., Fryer, C.~L.,
858: Woosley, S.~E., Langer, N., \& Hartmann, D.~H.\ 2003, \apj, 591, 288
859:
860: \bibitem[Heger \& Woosley(2002)]{heg2002} Heger, A., \&
861: Woosley, S.~E.\ 2002, \apj, 567, 532
862:
863: \bibitem[Inoue et al.(2005)]{ino2005} Inoue, S., Aoki, W.,
864: Suzuki, T.~K., Kawanomoto, S., Garc{\'{\i}}a-P{\'e}rez, A.~E., Ryan, S.~G.,
865: \& Chiba, M.\ 2005, in IAU Symp. 228, From Lithium to Uranium: Elemental Tracers of Early
866: Cosmic Evolution, ed. V.~Hill, P.~Fran{\c c}ois, \& F.~Primas (Cambridge: Cambridge University Press), 59
867:
868: \bibitem[Jedamzik(2000)]{jed2000} Jedamzik, K.\ 2000, \prl, 84, 3248
869:
870: \bibitem[Jedamzik(2004a)]{jed2004a} Jedamzik, K.\ 2004, \prd, 70,
871: 063524
872:
873: \bibitem[Jedamzik(2004b)]{jed2004b} Jedamzik, K.\ 2004, \prd, 70,
874: 08351
875:
876: \bibitem[Jedamzik(2006)]{jed2006} Jedamzik, K.\ 2006, \prd, 74,
877: 103509
878:
879: \bibitem[Jenkins et al.(2001)]{jen2001} Jenkins, A., Frenk,
880: C.~S., White, S.~D.~M., Colberg, J.~M., Cole, S., Evrard, A.~E., Couchman,
881: H.~M.~P., \& Yoshida, N.\ 2001, \mnras, 321, 372
882:
883: \bibitem[Kawano(1992)]{kawano1992} Kawano, L.\ 1992, NASA
884: STI/Recon Technical Report N, 92, 25163
885:
886: \bibitem[Kawasaki et al.(2005)]{kawasaki2005} Kawasaki, M., Kohri,
887: K., \& Moroi, T.\ 2005, \prd, 71, 083502
888:
889: \bibitem[Kneller et al.(2003)]{kne2003} Kneller, J.~P.,
890: Phillips, J.~R., \& Walker, T.~P.\ 2003, \apj, 589, 217
891:
892: \bibitem[Kusakabe et al.(2006)]{kus2006} Kusakabe, M., Kajino,
893: T., \& Mathews, G.~J.\ 2006, \prd, 74, 023526
894:
895: \bibitem[Letaw et al.(1983)]{let1983} Letaw, J.~R., Silberberg,
896: R., \& Tsao, C.~H.\ 1983, \apjs, 51, 271
897:
898: \bibitem[Maeder \& Meynet(1989)]{mae1989} Maeder, A., \&
899: Meynet, G.\ 1989, \aap, 210, 155
900:
901: \bibitem[Mathews et al.(2005)]{mat2005} Mathews, G.~J., Kajino,
902: T., \& Shima, T.\ 2005, \prd, 71, 021302
903:
904: \bibitem[Mel{\'e}ndez \& Ram{\'{\i}}rez(2004)]{mel2004}
905: Mel{\'e}ndez, J., \& Ram{\'{\i}}rez, I.\ 2004, \apjl, 615, L33
906:
907: \bibitem[Meneguzzi et al.(1971)]{men1971} Meneguzzi, M.,
908: Audouze, J., \& Reeves, H.\ 1971, \aap, 15, 337
909:
910: \bibitem[Mercer et al.(2001)]{mer2001} Mercer, D.~J., et al.\
911: 2001, \prc, 63, 065805
912:
913: \bibitem[Montmerle(1977)]{mon1977} Montmerle, T.\ 1977, \apj,
914: 216, 177
915:
916: \bibitem[Piau et al.(2006)]{pia2006} Piau, L., Beers, T.~C.,
917: Balsara, D.~S., Sivarani, T., Truran, J.~W.,
918: \& Ferguson, J.~W.\ 2006, \apj, 653, 300
919:
920: \bibitem[Pospelov(2006)]{pos2006} Pospelov, M.\ 2006, preprint
921: (hep-ph/0605215)
922:
923: \bibitem[Prantzos(2006)]{pra2006} Prantzos, N.\ 2006, \aap,
924: 448, 665
925:
926: \bibitem[Press \& Schechter(1974)]{pre1974} Press, W.~H., \&
927: Schechter, P.\ 1974, \apj, 187, 425
928:
929: \bibitem[Primas et al.(2000a)]{pri2000a} Primas, F., Asplund, M.,
930: Nissen, P.~E., \& Hill, V.\ 2000, \aap, 364, L42
931:
932: \bibitem[Primas et al.(2000b)]{pri2000b} Primas, F., Molaro, P.,
933: Bonifacio, P., \& Hill, V.\ 2000, \aap, 362, 666
934:
935: \bibitem[Primas et al.(1999)]{pri1999} Primas, F., Duncan,
936: D.~K., Peterson, R.~C., \& Thorburn, J.~A.\ 1999, \aap, 343, 545
937:
938: \bibitem[Ramaty et al.(1997)]{ram1997} Ramaty, R., Kozlovsky,
939: B., Lingenfelter, R.~E., \& Reeves, H.\ 1997, \apj, 488, 730
940:
941: \bibitem[Read \& Viola(1984)]{rea1984} Read, S.~M., \& Viola,
942: V.~E., Jr.\ 1984, Atomic Data and Nuclear Data Tables, 31, 359
943:
944: \bibitem[Reeves(1974)]{ree1974} Reeves, R.\ 1974, \araa, 12,
945: 437
946:
947: \bibitem[Ryan et al.(2000)]{rya2000} Ryan, S.~G., Beers, T.~C.,
948: Olive, K.~A., Fields, B.~D., \& Norris, J.~E.\ 2000, \apjl, 530, L57
949:
950: \bibitem[Rollinde et al.(2005)]{rol2005} Rollinde, E.,
951: Vangioni, E., \& Olive, K.\ 2005, \apj, 627, 666
952:
953: \bibitem[Rollinde et al.(2006)]{rol2006} Rollinde, E.,
954: Vangioni, E., \& Olive, K.~A.\ 2006, \apj, 651, 658
955:
956: \bibitem[Rollinde et al.(2008)]{rol2008} Rollinde, E., Maurin,
957: D., Vangioni, E., Olive, K.~A., \& Inoue, S.\ 2008, \apj, 673, 676
958:
959: \bibitem[Schaerer(2002)]{sch2002} Schaerer, D.\ 2002, \aap,
960: 382, 28
961:
962: \bibitem[Sheth \& Tormen(1999)]{she1999} Sheth, R.~K., \&
963: Tormen, G.\ 1999, \mnras, 308, 119
964:
965: \bibitem[Shi et al.(2007)]{shi2007} Shi, J.~R., Gehren, T., Zhang, H.~W.,
966: Zeng, J.~L., \& Zhao, G.\ 2007, \aap, 465,
967: 587
968:
969: \bibitem[Spergel et al.(2003)]{spe2003} Spergel, D.~N., et al.\
970: 2003, \apjs, 148, 175
971:
972: \bibitem[Spergel et al.(2007)]{spe2007} Spergel, D.~N., et al.\
973: 2007, \apjs, 170, 377
974:
975: \bibitem[Spite \& Spite(1982)]{spi1982} Spite, F., \& Spite,
976: M.\ 1982, \aap, 115, 357
977:
978: \bibitem[Suzuki \& Inoue(2002)]{suz2002} Suzuki, T.~K., \&
979: Inoue, S.\ 2002, \apj, 573, 168
980:
981: \bibitem[Suzuki \& Yoshii(2001)]{suz2001} Suzuki, T.~K., \&
982: Yoshii, Y.\ 2001, \apj, 549, 303
983:
984: \bibitem[Tatischeff \& Thibaud(2007)]{tat2007} Tatischeff, V.,
985: \& Thibaud, J.-P.\ 2007, \aap, 469, 265
986:
987: \bibitem[Valle et al.(2002)]{val2002} Valle, G., Ferrini, F.,
988: Galli, D., \& Shore, S.~N.\ 2002, \apj, 566, 252
989:
990: \end{thebibliography}
991:
992:
993: \begin{figure}
994: \begin{center}
995: \includegraphics[angle=-90,scale=0.5]{f1.eps}
996: \end{center}
997: \caption{Abundances of light elements in the IGM as a function of redshift
998: in Model 1 (solid lines). $\epsilon=0.31$ is assumed to result in
999: $^6$Li/H=$6\times 10^{-12}$ at $z=3$. The contribution of the normal
1000: mode stars only to the light element production is shown by the dashed
1001: lines.\label{fig1}}
1002: \end{figure}
1003:
1004: \begin{figure}
1005: \begin{center}
1006: \includegraphics[angle=-90,scale=0.5]{f2.eps}
1007: \end{center}
1008: \caption{Abundances of light elements in the IGM as a function of redshift
1009: in the rapid burst model. $\epsilon=0.029$ is assumed to result in
1010: $^6$Li/H=$6\times 10^{-12}$ at $z=3$.\label{fig2}}
1011: \end{figure}
1012:
1013: \begin{figure}
1014: \begin{center}
1015: \includegraphics[angle=-90,scale=0.5]{f3.eps}
1016: \end{center}
1017: \caption{Yields of light elements at $z=3$ by CCRs generated at $z_s$ per second
1018: in Model 1.\label{fig3}}
1019: \end{figure}
1020:
1021: \begin{figure}
1022: \begin{center}
1023: \includegraphics[angle=-90,scale=0.5]{f4.eps}
1024: \end{center}
1025: \caption{Plateau abundances of lithium isotopes produced by the CCR
1026: nucleosynthesis in Model 1
1027: with the accelerating efficiency $\epsilon=0.31$. $^7$Li data are from
1028: \citet[filled triangles]{asp2006}, \citet[open squares]{bon2007}, and
1029: \citet[open stars]{shi2007}. $^6$Li data are from \citet[large filled circles to detections, small filled circles
1030: to upper limits]{asp2006} and \citet[open circles to detections,
1031: crosses to upper limits]{ino2005}.\label{fig4}}
1032: \end{figure}
1033:
1034: \begin{figure}
1035: \begin{center}
1036: \includegraphics[angle=-90,scale=0.5]{f5.eps}
1037: \end{center}
1038: \caption{Plateau abundances of Be and B produced by the CCR nucleosynthesis
1039: in Model 1 (solid lines) with the accelerating efficiency
1040: $\epsilon=0.31$. The case that the normal mode star formation alone is
1041: considered as the CR energy source corresponds to the dashed lines with
1042: $\epsilon=0.73$ to realize the MPHS value of $^6$Li at $z=3$. Plateau abundances in the rapid burst model with the accelerating efficiency
1043: $\epsilon=0.029$ are also shown as the dotted lines. $^9$Be data are
1044: from \citet[filled circles]{boe1999}, \citet[open circles]{pri2000a},
1045: \citet[open triangle]{pri2000b}, and \citet[open stars]{boe2006}. B
1046: data are from \citet[filled squares]{dun1997}, \citet[open
1047: squares]{gar1998}, \citet[filled triangles]{pri1999}, and
1048: \citet[crosses]{cun2000}.\label{fig5}}
1049: \end{figure}
1050:
1051: \begin{figure}
1052: \begin{center}
1053: \includegraphics[angle=-90,scale=0.5]{f6.eps}
1054: \end{center}
1055: \caption{Energy fraction of SNRs to CRs, $\epsilon$, as a function of the
1056: index of the CR injection spectrum, $\gamma$, required to produce
1057: $^6$Li at the MPHSs level $^6$Li/H=$6\times 10^{-12}$ at $z=3$. The
1058: solid line corresponds to Model 1 and the dashed line to the case where
1059: only the normal mode stars are considered as the energy source of
1060: CCRs.\label{fig6}}
1061: \end{figure}
1062:
1063: \begin{figure}
1064: \begin{center}
1065: \includegraphics[angle=-90,scale=0.5]{f7.eps}
1066: \end{center}
1067: \caption{Abundances of Be and B at $z$=3 in Model 1 with the CR energy
1068: fraction $\epsilon$ in Fig.\ \ref{fig6}, when $^6$Li is produced at the
1069: MPHSs level. The solid lines correspond to Model 1 and the dashed lines
1070: to the case where only the normal mode stars are considered as the energy
1071: source of CCRs.\label{fig7}}
1072: \end{figure}
1073:
1074: \begin{figure}
1075: \begin{center}
1076: \includegraphics[angle=-90,scale=0.5]{f8.eps}
1077: \end{center}
1078: \caption{Abundances of light elements produced by the secondary process
1079: in the ISM as a function of redshift in Model 1 (solid
1080: lines). $\epsilon=0.31$ is assumed to result in $^6$Li/H=$6\times 10^{-12}$ in the IGM at $z=3$. The contribution of the normal mode stars only to the light element production is shown by the dashed lines.\label{fig8}}
1081: \end{figure}
1082:
1083: \clearpage
1084:
1085: \begin{deluxetable}{ccccccc}
1086: \tablecaption{Abundance results for two models\label{tab1}}
1087: \tablewidth{0pt}
1088: \tablehead{
1089: \colhead{Model} & \colhead{$\epsilon$} & \colhead{$^{6}$Li/H} &
1090: \colhead{$^{7}$Li/H} & \colhead{$^{9}$Be/H} & \colhead{$^{10}$B/H} &
1091: \colhead{$^{11}$B/H}
1092: }
1093: \startdata
1094: Model 1 & 0.31 & 6.0$\times~10^{-12}$ & 7.3$\times~10^{-12}$ & 9.0$\times~10^{-14}$ & 3.2$\times~10^{-13}$ & 7.6$\times~10^{-13}$ \\
1095: Model 1 (Pop II only)& 0.73 & 6.0$\times~10^{-12}$ & 7.3$\times~10^{-12}$ & 5.7$\times~10^{-14}$ & 2.1$\times~10^{-13}$ & 4.9$\times~10^{-13}$ \\
1096: Rapid burst model & 0.029 & 6.0$\times~10^{-12}$ & 7.9$\times~10^{-12}$ & 4.4$\times~10^{-13}$ & 2.0$\times~10^{-12}$ & 5.2$\times~10^{-12}$ \\
1097: \enddata
1098:
1099: \end{deluxetable}
1100:
1101:
1102: \end{document}
1103:
1104:
1105:
1106:
1107: