Origin of matter and evolution of galaxies 2000

Editors Taka Kajino "ihigeru Kubono

^en-ichi Nomoto

■ sao Tanihata

O r i g i n 0/ M a t t e r ^ E v o l u t i o n 0/ Galaxies 2

0

0

0

O r i g i n 0/ M a t t e r ^ E v o l u t i o n 0/ Galaxies 2

0

0

0

This page is intentionally left blank

Editors Taka Kajino National Astronomical

Observatory, Japan

Shigeru Kubono Center for Nuclear Study, University of Tokyo, Japan

Ken-ichi Nomoto Department of Astronomy, University of Tokyo, Japan

Isao Tanihata Rl Beam Science Laboratory, RIKEN, Japan

O r i g i n of M a t t e : & . E v o l u t i o n of G a l a x i e s 2

0

0

I© World Scientific ■

New Jersey • London • Singapore • Hong Kong

0

Published by World Scientific Publishing Co. Pte. Ltd. 5 Toh Tuck Link, Singapore 596224 USA office: Suite 202, 1060 Main Street, River Edge, NJ 07661 UK office: 57 Shelton Street, Covent Garden, London WC2H 9HE

British Library Cataloguing-in-Publication Data A catalogue record for this book is available from the British Library.

ORIGIN OF MATTER AND EVOLUTION OF GALAXIES 2000 Copyright © 2003 by World Scientific Publishing Co. Pte. Ltd. All rights reserved. This book, or parts thereof, may not be reproduced in any form or by any means, electronic or mechanical, including photocopying, recording or any information storage and retrieval system now known or to be invented, without written permission from the Publisher.

For photocopying of material in this volume, please pay a copying fee through the Copyright Clearance Center, Inc., 222 Rosewood Drive, Danvers, MA 01923, USA. In this case permission to photocopy is not required from the publisher.

ISBN 981-238-287-9

Printed in Singapore by World Scientific Printers (S) Pte Ltd

International Symposium on

Origin of Matter and Evolution of Galaxies 2000 Center for Nuclear Study, University of Tokyo Tanashi, Tokyo January 19 - 21, 2000

Organizing Committee of the Symposium: T. Kajino National Astronomical Observatory , co-chairman S. Kubono CNS, University of Tokyo, co-chairman T. Kifune ICRR, University of Tokyo K. Sato University of Tokyo Y. Suzuki ICRR, University of Tokyo I. Tanihata RIKEN H. Toki RCNP, Osaka University H. Miyatake KEK T. Motobayashi Rikkyo University Y. Nagai RCNP, Osaka University K. Nomoto University of Tokyo Hosted by:

Center for Nuclear Study, University of Tokyo Division of Theoretical Astrophysics, National Astronomical Observatory RI Beam Science Laboratory, RIKEN E-Group, KEK

V

This page is intentionally left blank

Preface This book is the proceedings of the International Symposium on Origin of Matter and Evolution of Galaxies which was held in Tokyo, Japan, during January 19 21 in 2000. We have organized a series of international meetings on this subject several times since 1992, and this meeting was held in a special occasion twofold. First, we are in the first year of new millenium that should bring us to a new stage of science, based on largely accumulated scientific knowledge and highly developed technology in the last century. Second, this symposium was the very last meeting held in Tanashi campus of the Center for Nuclear Studies, University of Tokyo, which has moved to Wako campus. The purpose of this symposium was to bringmany scientists together from vast science fields, i.e. nuclear physics, particle physics, cosmic-ray physics, cosmology, astronomy, geophysics, and others, in order to promote cooperative discussions and collaboration. We have made several scientific progress in recent years in this interdescipUnary field named Nuclear Cosmology and Astrophysics. Radioactive IonBeam Facilities at many institutes such as RTKEN, CNS, KEK, GSI, MSU, Oak Ridge, and others succeeded in producing (or planning to produce) new isotopes to study nuclear structure and reactions of exotic nuclei, and these new nuclei have quickly been used secondarily for the studies of nuclear astrophysics. Explosive nucleosynthesis in the Big-Bang Supernovae or Novae are now being deeply investigated both experimentally and theoretically, and the world biggest Telescopes SUBARU, KECK, and VLT will start operation to provide valuable and complemantary informatoion on understanding the origin and evolution of elements in the Universe. Kamiokande op ened anew frontier of neutrino astronomy, and Super-Kamiokande is expected to reveal the nature of neutrino. In addition to these terrestrial experimental projects, science has been done from celestial laboratories as well. X-ray satellite ASCA has revealed many interesting asp ects of activites of neutron star, black hole, extra-galaxies and cluster of galaxies. Radio-wave satellite HALCA was successfully launched and has started operation. All these are now being carried out in international collaborations. We actually live in an exciting epoch at the beginnig of the 21 s t century. It was really a great pleasure to organize this high quality international symposium in Japan. Taka Kajino, Co-Chairman National Astronomical Observatory Shigeru Kubono, Co-Chairman Center for Nuclear Study, University of Tokyo Ken-ichi Nomoto Department of Astronomy, University of Tokyo Isao Tanihata RI Beam Science Laboratory, RJKEN

VII

This page is intentionally left blank

Opening address

Dear Colleagues, good morning !

I would like to make a few words before starting the session. Today, we are very happy to have you all to this international symposium on Origin of Matter and Evolution of Galaxies 2000. We would like to thank all the participants, especially who came from outside of Japan for a long distance. In fact, this is the third meeting under the same title. Our primary interest is to get together the astronomers and physicists to discuss common problems of the Universe. Especially, here, we are emphasizing the importance of an axis of not only the elements but also the isotopes, namely the atomic nuclides, which should give us an important clue for understanding the phenomena of the Universe. This is a good occasion to make a scope of the field at the beginning of the new millennium. Our intention is to have, not only presentation of the completed works, but also to discuss the current problems and the possible future collaborations. We hope that you all enjoy the active and stimulating discussions in this symposium.

Thank you. S. Kubono Chairperson of the Organizing Committee

IX

This page is intentionally left blank

Contents

Preface Opening Address S. Kubono (Co-Chairperson,

vii ix CNS)

I. Early Universe and Chemo-Dynamic Evolution of Galaxies T. Kajino Prospects in Nuclear Cosmology: From the Big-Bang to Supernovae

3

T. Shigeyama Inhomogeneous Chemical Evolution in the Galactic Halo: Supernova-Induced Formation of Field Stars and Globular Clusters

14

T. Suzuki Light Elements in Inhomogeneous Early Galaxy and Their Astrophysical Interests

23

T. Kifune Prospects for Very High Energy 7-Ray Astronomy with Next Generation Imaging Cerenkov Telescope

31

H. Umeda Evolution and Explosion of Massive Pop III Stars and Their Nucleosynthesis

43

II. Observation of Elements S. Amari Presolar Grains as Probes of Nucleosynthesis in Stars and Evolution of the Galaxy

XI

55

XII

C. Iliadis Nucleosynthesis of Mg and Al in Globular Cluster Red Giant Stars

69

Y. Fukazawa X-Ray Measurements of Metal Abundances of Hot Gas in Clusters of Galaxies

77

H. Murakami X-Ray Diagnosis of the Galactic Center Abundance with an X-Ray Reflection Nebula

86

N. Hasebe Cosmic Ray Observation for Nuclear Astrophysics: CORONA Program

94

III. Stellar Evolution and the Nucleosynthesis: Hydrostatic Burning R.E. Tribble Direct Capture S-Factors from Asymptotic Normalization Coefficients

107

W.-P. Liu Solar Neutrino Problem Related Nuclear Physics Experiments

119

N. Iwasa Coulomb Dissociation of 8 B at 254 MeV/u for 7 Be(p, 7) 8 B

130

N. Kudomi Development for the Study of a Cross-Sectional Measurement of 3 He- 3 He Solar Reaction

138

IV. Nucleosynthesis in Explosive Burning and N e w Approach M.S. Smith Probing Stellar Explosions with Radioactive Beams at ORNL

149

J. D'Auria Measuring the Astrophysics Rate for Radiative Proton Capture on 21 Na

163

XIII

S. Kubono Nuclear Astrophysics Project with a New Low-Energy RIB Separator CRIB: Study of a Critical Stellar Reaction 15 0(a,7)19Ne

171

V. Explosion of Massive Stars S. Yamada Physics of Collapse-Driven Supernovae

181

M. Yasuhira Protoneutron Stars with Kaon Condensate and Possibility of Delayed Collapse

194

R.N. Boyd OMNIS, the Observatory for Multiflavor Neutrinos from Supernovae

201

Y. Fukuda Observation of Supernova Neutrino Burst at Super-Kamiokande

209

K. Homma Can the Negative Mass Square of the Electron Neutrinos Be an Indication of Interaction with Relic Neutrinos?

215

K. Nomoto Nucleosynthesis in Hypernovae

223

H. Tsunemi Overabundance of Calcium in the Young SNR RX J0852-0462: Evidence of Over-Production of 44 Ti

240

K. Koyama X-Ray Spectroscopy and Chemical Composition in the Universe

246

VI. Origin of Heavy Elements G.J. Mathews Neutron Star Mysteries

257

xiv H. Utsunomiya Photoneutron Cross Sections for 9 Be and the a-Process in Core-Collapse Supernovae

267

T. Suda RIKEN RI-Beam Factory (RIBF) Project and the Way to the r-Process Nuclei

276

M. Hashimoto Connection Between Crucial Nuclear Reaction Rates and the Modeling of Accreting Neutron Stars

283

VII. Neutron Stars and High Density Matter K. Sumiyoshi Unstable Nuclei and an EOS Table for Supernovae and the r-Process in a Relativistic Many-Body Approach

297

T. Takatsuka Baryon Superfluidity in Neutron Star Cores

305

T. Tatsumi Ferromagnetism of Quark Liquid and Magnetars

313

T. Kishimoto Kaonic Nuclei and Kaon Condensation in Neutron Stars

322

Poster Session A. Bamba Chemical Composition and Distribution of Heavy Elements in a Supernova Remnant G359.1-0.5

331

H. Ishiyama Tanashi Recoil Mass Separator for Nuclear Astrophysics

335

A. Iwazaki Collision Between Neutron Star and Axion Star as a Source of Gamma Ray Burst and Ultra-High Energy Cosmic Ray

339

XV

N. Kudomi Double Beta Decays of Observatory

100

Mo by ELEGANT V at Oto Cosmo 343

K. Maeda Nucleosynthesis in Aspherical Hypernova Explosions and Late Time Spectra of SN 1998BW

347

S. Nagataki Effects of Jet-Like Explosion in SN 1987A

354

N. Nakasato Formation and Chemical Dynamics of the Galaxy

358

J. Nakatsuru Explosive Nucleosynthesis in Pair-Instability Supernovae

362

M. Nishiuchi X-Ray Observations of SNRs and Hot ISM in the Large Magellanic Cloud: The Chemical Enrichment of the Galaxy

370

K. Otsuki r-Process Nucleosynthesis in Neutrino-Driven Wind: General Relativistic Effects and Short Dynamic Timescale Model

374

M. Terasawa The Critical Role of Light Neutron-Rich Nuclei in the r-Process Nucleosynthesis

378

G. Watanabe Thermodynamic Properties of Nuclear "Pasta" in Neutron Star Crusts

381

Symposium Program

385

List of Participants

391

Author Index

403

This page is intentionally left blank

I. Early Universe and Chemo-Dynamic Evolution of Galaxies

This page is intentionally left blank

Prospects in Nuclear Cosmology: From t h e Big-Bang to Supernovae Toshitaka Kajino National Astronomical Observatory Mitaka, Tokyo 181-8588 and Department of Astronomy, University of Tokyo Bunkyo-ku, Tokyo 113-0033

Abstract We study primordial nucleosynthesis in the presence of a net lept.on asymmetry. We explore a previously unnoted region of the parameter space in which very large baryon densities 0.1 < fit, < 1 can be accommodated within the light-element constraints. This parameter space consists of large v,, and uT degeneracies with a moderate ve degeneracy. Constraints on this parameter space from cosmic microwave background fluctuations are discussed [1]. We also study the r-process nucleosynthesis in neutrino-driven winds of gravitational core collapse SNell. Appropriate physical conditions are found for successful r-process nucleosynthesis, which meet, with several features of heavy elements discovered recently in metal-deficient, halo stars.

1

Big-Bang Cosmology

Recent progress in cosmological deep survey has clarified progressively the origin and distribution of matter and evolution of Galaxies in t.he Universe. The origin of the light elements among them has been a topic of broad interest, for its significance in constraining the dark matter component in the Universe and also in seeking for the cosmological model which best, fits the recent, data of cosmic microwave background (CMB) fluctuations. This paper is concerned with neutrinos during Big-Bang nucleosynthesis (BBN). In particular, we consider new insights into the possible role which degenerate neutrinos may have played in the early Universe. There have been many important contributions toward constrainig neutrino physics. Hence, a discussion of neutrinos and BBN is even essential in particle physics as well as cosmology. There is no observational reason to insist that the universal lept.on number is zero. It is possible, for example, for the individual lept.on numbers to be large

3

4

compared to the baryon number of the Universe, while the net total lepton number is small L ~ B. It has been proposed recently [2] that models based upon the Affleck-Dine scenario of baryogenesis might generate naturally lepton number asymmetry which is seven to ten orders of magnitude larger than the baryon number asymmetry. Neutrinos with large lepton asymmetry and masses ~ 0.07 eV might even explain the existence of cosmic rays with energies in excess of the Greisen-Zatsepin-Kuzmin cutoff [?>]. It. is, therefore, important for both particle physics and cosmology to carefully scrutinize the limits which cosmology places on the allowed range of both the lepton and baryon asymmetries.

1.1

Cosmological Neutrino and Primordial Nucleosynthe­ sis

Although lepton asymmetric BBN has been studied in many papers [4] (and references therein), there are several differences in the present, work: For one , we have included finite temperature corrections to the mass of the electron and photon [5]. Another is t h a t we have calculated the neutrino annihilation rate in the cosmic comoving frame, in which the M0ller velocity instead of the relative velocity is to be used for the integration of the collision term in the Boltzmann equations [6, 7]. Neutrinos and anti-neutrinos drop out of thermal equilibrium with the back­ ground thermal plasma when the weak reaction rate becomes slower than the universal expansion rate. If the neutrinos decouple early, they are not heated as the particle degrees of freedom change. Hence, the ratio of the neutrino to photon temperatures, Tu/Xy, is reduced. The biggest drop in temperature, for all three neutrino flavors occurs for £„ ~ 10. This corresponds to a decoupling temperature above the cosmic QCD phase transition. Non-zero lepton numbers affect, nucleosynthesis in two ways. First, neutrino degeneracy increases the expansion rate. This increases the ''He production. Secondly, the equilibrium n / p ratio is affected by the electron neutrino chemical potential, n / p = exp{ — (AM /Tnu ) — &,,.}, where AM is the neutron-proton mass difference and T„ 0.4 only the large degeneracy solution is allowed. Neutrino degen­ eracy can even allow baryonic densities up to Qbh^Q = 1.

5 1,4 T

n

,.''i=L0

1.2-

..--■'

..

"

.•■'

-i

]

mi- =o.3

o.8-

■;'

;/.'. ■

-.

]

«- k f 11.2 - :".

■S-iiJi:

=0.075

l),»r'. ■ . . ■ . . . . II

l . . . . III

\ i . . . .

i .. 20

I 3d

:

4ll

50

Figure 1: Allowed values of £,,c and £„, r for which the constraints from light. element abundances are satisfied for values of Ab^lo — 0-075, 0.1, 0.2, 0.3 and 1.0 as indicated.

1.2

Cosmic Microwave Background

Several recent works [8, 9, 10] have shown that neutrino degeneracy can dramat­ ically alter the power spectrum of the CMB. However, only small degeneracy parameters with the standard relic neutrino temperatures have been utilized. Here, we have calculated the CMB power spectrum to investigate effects of a diminished relic neutrino temperature. The solid line on Figure 2 shows a QA = 0.4 model for which ??. = 0.78. This fit is marginally consistent with the data, at a level of 5.2-

o

0

20

40

60

80

100

120

140

160

180

200

220

240

MASS NUMBER Figure 3: R-process abundance [20] (solid line) as a function of atomic mass number A compared with the solar system r-process abundance (filled circles) from Kappeler, Beer, k Wisshak [23]. The neutrino-driven wind model used is for L„ = 1052 ergs/s and M = 2M?,. The solar system r-process abundance is shown in arbitrary unit.

Let us remind the readers that, there were at least three difficulties in the previous theoretical studies of the r-process. The first difficulty among them is that an ideal, high entropy in the bubble S/k ~ 400 [16] is hard to be achieved in the other simulations [17, 18, 19, 20]. The key to resolve this difficulty is found with the short dynamic time scale Tdyn ~ 10 ms in our models of the neutrino-driven winds. As the initial nuclear composition of the relativisfic plasma consists of neutrons and protons, the oburning begins when the plasma temperature cools below T ~ 0.5 MeV. The 4 ]-le(aa,'))1-C reaction is too slow at this temperature, and alternative nuclear reaction path ' 1 He(an, -))f,Be(a, ?7)12C triggers explosive a-burning to produce seed elements with A ~ 100. Therefore, the time scale for nuclear reactions is regulated by the 4 He(a?i,7)''Be. It is given by T^ ~ (piY£Y„\(aan -¥w Be)) If the neutrino-driven winds fulfill the condition Tcly„ < 7w, then fewer seed nuclei are produced during the a-process with plenty of free neutrons left over when the r-process begins at T ~ 0.2 MeV. The high neufron-to-seed ratio, n/s ~ 100, leads to appreciable production of r-process elements, even for low entropy S/k ~ 130, producing both the 2nd (.4 ~ 130) and 3rd (.4 ~ 195) abundance peaks and the hill of rare-earth elements (.4 ~ 165) (Figure 3).

10

100

150

200

mass number

Figure A: The same as (hose in Figure 3, but for the neutrino-driven wind model of Lv — 5 x 1052 ergs/s. Solid line respresents the result, by using the Woosley fe Hoffman rate [25] of the 4He(an,7)''Be reaction, and long-dashed line for the rate multiplied by factor 2, as suggested by the recent experiment of Utsunomiya et al. [2(3]. The three body nuclear reaction cross section for *1He(o», 7)!'Be is one of the poorly determined nuclear data which may alter the r-process nucleosyn­ thesis yields. The inverse process has recently been studied experimentally by Utsunomiya et al. [2(3], and phofodisinfegration cross section of ;'Be has been measured with better precision than those of the previous experiments. Ap­ plying the principle of the detailed balance to this process, one can estimate the cross section for 4Ue(an,7)9Be. They found that the thermonuclear reac­ tion rate is almost, twice as big as that of Woosley and Hoffman [25] but in resonable agreement with recent compilation of Angulo et al. [27]. However, there still remain several questions on the consistency between their result and electron-scattering experiments, on the contribution from the narrow resonance J* = 5/2 _ (2.429 MeV), etc. It is also a theoretical challenge to understand the reaction mechanism and the resonance structure because two different channels, 8 Be + n and °He + a, contribute to this process. Therefore, we show two calculated results in Figure 4: The solid line displays the result, obtained by using the Woosley and Hoffman cross section [25], assum­ ing a 8 Be -f- n structure for !'Be. We also calculated the r-process by multiplying this cross section by factor of 2 (long-dashed line). This makes a drastic change in the r-process yields in the 3rd (A ~ 195) abundance peak. More theoretical and experimental studies of the '1He(o7i,7)!'Be reaction are highly desired.

11

2.3

Neutrino-nucleus interactions

Neutrino interactions with nucleons and nuclei take the key to resolve the second difficulty which was pointed out in sect. 1. The difficulty is that the effects of neutrino absorptions vt + n —> p + c~~ and ve + A(Z,N) —> A{Z + 1, N — 1) + e~ during the a-process may induce the deficiency of free neutrons and break down the r-process conditions [21]. These two types of neutrino interactions control most sensitively the electron fraction and the neutron fraction, as well, in a neutron-rich environment. In order to resolve this difficulty, we have updated the electron-type neutrino capture rates for all nuclei and electron-type antineutrino capture rate for free proton [28, 29]. The new r-process calculation proves to be almost invariant. One can un­ derstand this robustness of the succesful r-process in the following way: The specific collision time for neutrino-nucleus interactions is given by

^20lx [ ; i ,x(^)( T ^) 2 ( B M i ? )"',, > ,

„)

where Lj^i is the individual neutrino or antineufrino luminosity in unit.s of It)51 ergs/s, (j =< E'f > I < Ej > in MeV (/ = ve, />t, etc.), and (cr„) is the averaged cross section over neutrino energy spectrum. At the n-burning site of r « 100 km for L„i5i « 10, (m) used here is a Salpeter one with a slope irdex of —1.35. The upper mass limit of stars is assumed to be mu = 50 MQ, and the lower mass limitTO;= 0.05 MQ 15 . The lower mass limit of stars that explode as SNe is TOSN,; = 10MQ. Though the form of the function F introduced in equation (1) is not known, it can be expanded into a Taylor series around the point where no SN happens: Af,( 10 4 m 2 , almost five orders of magnitude larger than the satellite instrumentation at MeV to GeV energies. The imaging technique of Cerenkov lights is capable of rejecting about 99% of the overwhelming background of cosmic ray proton events. These conditions could bring about the "break-through" of detection technique; by firmly confirming the VHE 7-ray signal from the Crab nebula, thanks to the fact that the Crab flux of ~ 1 0 - 1 1 c m _ 2 s _ 1 (E 7 > 1 TeV) was just as intense as the sensitivity of the imaging Cerenkov telescope with observation time of tens of hours. The observation of the past 10 years has shown that further improvements are necessary to go beyond the limitation the current results have, for instance, in the number of discovered objects and in the distance to the sources that detection sensitivity allows. The world-wide efforts are thus attempting to construct the "next generation telescopes", which will reduce the detectable energy of 7-rays down to ~100 GeV. Observation will start in a few years to in more details uncover violent, energetic phenomena in which high energy acceleration of electrons and protons plays a dominant role. 2

High energy 7-ray sources

Considerable efforts in 7-ray observation have been spent in search for exotic phenomena, such as antimatter or primordial black holes, that could inform us of the debris of high energy particle interaction in earlier stages of the Universe. The searches have so far betrayed such speculative expectation. Instead, a wealth of non-thermal processes have been unexpectedly uncovered, as their typical energy spectra illustrated in Fig. 1, with a considerable number of high energy 7-ray sources as listed in Table 1 from the third catalogue (Hartman et al. 1999) of EGRET detection of CGRO (Compton Gamma Ray Observatory). The case of VHE 7-rays is also added in the bracket. VHE 7-rays are detected so far from AGN (active galactic nuclei), pulsar nebula, SNR (supernova remnant) and possibly from an X-ray close binary, Cen X-3 (Vestrand et al. 1997; Chadwick et al. 1998). We may indicate that there exist interesting dissimilarities between the two energy regions: There is no clear suggestion of stronger GeV sources likely to be a latent TeV source. Two distinct classes appear to exist in blazars, i.e. 7-ray emitting AGN, of GeV and TeV ones. The TeV blazars are not bright in GeV 7-rays and, rather, belong to less luminous sources of the EGRET catalogue (Ghisellini et al. 1998). Emission of GeV 7-rays is from the pulsar magnetosphere while VHE 7-rays are from pulsar nebulae. Conversely, pulsar nebulae are not detected as GeV source except for the Crab nebula. Such a transition of the features of 7-ray source appear to take place in the energy region of 10 -100 GeV, which remains

33

-7

i

GRB

_

-

% ° -9

Visible Star

^CrabNebula

e? o PL,

o

--1

-

f\^AGN

\

-11

—

-13

ys^

/ 1

\

SNR

s *' */ Galactic Disc " \ v Extra-galactic Diffuse" s

U

L

leV

1 keV

1 1 MeV

1 1 GeV

V

N 1 TeV

Photon Energy Figure 1: Schematic sketch of energy spectrum from 7-ray sources.

unexploited above the satellite and below the ground-based observation. The next generation of ground-based observation aims at this energy region, where a larger number of sources than in the TeV band are expected. The particle nature of photons becomes more prominent with increasing photon energy. Energetic 7-rays have inreraction of creating electron-positron pairs when they encounter photons of ambient softer radiation. The interacton sets limitation to the travelling distance to be no larger than ~ 100 Mpc for ~ 1 TeV 7-rays. The source must be optically thin for 7-rays to escape out of emission region. Weaker and remoter sources are likely to be disclosed by reducing the detectable energy of 7-rays. 3

Overview of the results in TeV energy region

The main part of VHE 7-ray astronomy remains still hidden below the current detection sensitivity. For instances, the emission from the Galactic disc is the most intense one in the GeV energy region and should extend to TeV energies, however, not yet detected at TeV energies because of the difficulty to detect extended emission with the current imaging Cerenkov technique. The imaging

34 Table 1: GeV (TeV) 7-ray sources

type of sources normal galaxy active galaxy

pulsar

unidenfified sources galactic latitude |6| < 1 0 ° galactic latitude |6| > 10°

number of GeV (TeV) sources 1 83 (2-4)

6(3)

note Large Magellanic Cloud quasar, BL Lac objects pulsed emission from pul­ sar magnetosphere in GeV energy (unpulsed from pulser nebula in TeV energy)

181 78

Violently time variable sources are included.

103

supernova remnant

?(2)

possibly sources

in

unidentified

X-ray binary

1(1?)

transient emission in GeV energy (Cen X-3)

Cerenkov telescope has a typical field of view of about 3°-5° diameter, and thus the survey over all the sky so far made is incomplete, particularly for time-variable, transient sources. 3.1 Extragalactic objects The active galaxies of GeV 7-ray emission distribute in red shift z as far as 2. However, the two firmly established ones of VHE 7-ray emission, Mrk 421 and Mrk 501 are of z~0.03. The distance that TeV 7-rays can travel through gets shorter with increasing 7-ray energy due to pair production of electron and positron in collision with softer photons with a greater spacial density of background radiation in extragalactic space. The maximum distance is estimated to be within about z~0.1 for 7-rays at ~1 TeV. The high energy activity of producing 7-rays are considered due to a mas­ sive black hole located in the central core of the active galaxies, i.e. AGN. The intense radiation field near the central engine of AGN was thought to prevent high energy 7-rays from escaping out. High energy 7-rays can come out against

35

such expectation thanks to the relativistic bulk motion of the AGN jet. The jet moving with a Lorentz factor T can enhance the energy and luminosity of 7-rays in the observer's frame than those in the rest frame of the jet by a factor T and T 4 ( r from each of the transformation of energy and time, and T2 from solid angle), respectively, thus reconciling the absorption effect due to electron-positron pair creation. Production of 7-rays is explained by energetic electrons scattering ambient soft photons up to 7-ray energies. By comparing this inverse Compton yield with synchrotron radiation into X-ray band, the 7-ray data could determine the parameters of the AGN jet such as magnetic filed B ~ 0.1G and T ~ 10. The data from VHE 7-rays are found useful to infer these values, but available objects are few to compare with those in the other bands. Another important question to be addressed and left for further efforts to answer is whether the energy of the jet is supplied by hadrons, i.e. due to acceleration of protons, or by non-hadronic process of electrons and positrons. In other words, there remains a question if AGN can be the source of extragalactic cosmic ray protons. 3.2

Galactic objects

CANGAROO (Collaboration between Australia and Nippon for a GAmma Ray Observatory in the Outback) has VHE 7-ray observation in Woomera, South Australia by using imaging Cerenkov telescope. The site has advantage of seeing the Galactic center near the zenith, and its main efforts were spent in surveying Galactic objects; mainly on pulsars and SNRs (e.g. Kifune 2000). The positive detections so far reported are from three pulsar nebulae of Crab, Vela, PSR1706-44 pulsars and from two SNRs of SN1006 and RXJ1713.7-3946 (Muraishi et al. 2000). In the case of Crab nebula, the observational data well cover wide energy bands to provide an "entire energy spectrum" of a two-peaked distribution which is interpreted as due to synchrotron and inverse Compton emissions and indicates that non-thermal electrons are the progenitor of the radiation. The spectra of the other sources also appear to be consistent with the view of electron progenitor. New results in observation generally stimulate new problems which re­ quire further efforts for advanced observation, in the case of VHE 7-rays from Galactic objects, for instances: (i) The two spectra of synchrotron and in­ verse Compton radiation may not be fully consistent with each other, or the entire spectrum needs to be known with better coverage over all the energy bands to enable the comparison with reasonably good accuracy, in order to confirm/clarify various kinds of different populations of progenitor particles; i.e. protons, fresh young or old aged electrons, (ii) Except for the case of

36

the Crab nebula, inverse Compton emission is considered to be from the col­ lision with 2.7K photons of cosmic microwave background, and the intensity ratio to synchrotron radiation suggests magnetic field as low as the value in the interstellar space ~ 3/uG. It is not clear what the low magnetic field im­ plies in relevance to particle acceleration mechanism, (iii) Emission region of VHE 7-rays from the Vela pulsar direction appears to be at the birth place of the pulsar which is displaced by about 0.3° from the present position. In order to have conclusive arguments about what these results may suggest, it is necesary to know the energy spectrum over broad, entire bands in more de­ tails for each source and on a greater number of Galactic VHE 7-ray sources than the current known ones; three pulsar nebulae, two SNRs and possibly one X-ray emitting close binary (Cen X-3). Electron progenitor is consistent with the data from pulsar nebulae and SNRs, but the currently known "entire" energy spectrum contain uncertainties which may allow, at least partially, the possibility of proton progenitor. 4

The next generation telescopes

As the next step of the ground-based telescope for VHE 7-rays, a system of multiple telescopes of 10m size is coming soon to provide a larger detection area added from many telescopes, better resolution of energy and arrival direction than single telescope. "Stereoscopic, simultaneous" operation of the multiple telescopes is intended to detect 7-rays in the energy region down to ~100 GeV. The concept is persued by the three projects. The CANGAROO group has started to construct four 10m telescopes, as a 5 years project, with a stereoscopic operation of two telescopes available in the 2001 year (Mori et al. 1999). The photograph of the first 10m telescope of the CANGAROO project is shown in Fig. 2. Two other projects, VERITAS collaboration in USA and H.E.S.S. of Euro­ pean collaboration lead by Max Planck Institute in Heidelberg, Germany, have proposed to construct 7 and 16 telescopes, respectively. The both projects ex­ pect to commence their first stage of constructing 4 telescopes in a few years almost at the same time with CANGAROO. The telescopes are installed with mutual distance of about 100 m which is less than the spread of the Cerenkov light pool to optimize detection area and rejection power of cosmic ray back­ ground. Slightly differently, another direction for the next generation of imag­ ing Cerenkov telescope is to achieve energy threshold of detectable 7-ray energy less than 50 GeV which overlaps the satellite detection, and the group lead by Max Planck Institute in Munich has started MAGIC project to construct a 17m aperture telescope.

37

Figure 2: The Cerenkov imaging telescope of 10m diameter of CANGAROO project.

The detectable energy of 7-rays is expected to be ~100 GeV by collect­ ing Cerenkov lights with the 10m aperture of reflection mirror. The largest telescope at present, the 10m telescope of Whipple group, has been, however, the threshold enegy of ~300 GeV, because Cerenkov lights emitted from single cosmic ray muon passing near the telescope produce a new kind of background events that resembles 7-ray signal of lower energies. A simple or easy way of distinguishing this background is to confirm no corresponding signal in the other telescopes of the stereoscopic operation, or improvements are necessary in the imaging camera of smaller pixel size as well as in the electronics circuit of better timing information. In addition to new sources that will be found in the energy region so far unexploited, the statistics of gathered number of 7-ray photons can be improved by going to lower 7-ray energies. The detection of energy flow at sensitivities better than 10 - 1 1 erg cm _ 2 s _ 1 is expected from observation duration time as short as 10 hrs for the sources that do not have very hard energy spectrum of VHE 7-rays. Comparison with the other energy bands of electromagnetic

38

radiation for the multi wavelength analysis will be made possible for a decent number of samples of VHE sources. The system of multiple telescopes will be operated in various ways. The opportunities of observing a number of targets are increased by setting each telescope to aim at each different objects. Alternatively, one same object is tracked simultaneously by multiple telescopes, and the effective duration time of observation or detection area is made larger. When multiple telescopes are within the area of the Cerenkov light pool of one event, "stereoscopic" image of 7-ray air shower event become available. Single imaging Cerenkov telescope has a power of identifying and rejecting about 99% of cosmic ray charged particles which is about 105 times more frequent than the intense flux from the Crab nebula. The rejection power can be increased in the stereoscopic observation as a product of the contributions from individual telescopes. The angular resolution of the single telescope depends on the direction relative to the elongation axis of Cerenkov light image. This drawback can be improved by using multiple telescopes and we will obtain angular resolution of ~ 0.1° for a single event. Thus, the stereoscopic observation makes it practically possible or easier to detect extended emission as well as weak sources so far hidden in cosmic ray background events. Search for more sources of AGN, pulsars and SNRs, is of prime importance in what will be done in the years to come. Our Galaxy can be surveyed for VHE sources brighter than the luminosity ~ 1032 erg s _ 1 within the distance as far as 10 kpc. A considerable number of 7-ray sources will be added to the multiwavelengths analysis. A systematic study of energy spectra from many AGN will determine the infrared intensity in extragalactic space and give estimate on the activities of galaxies, as well as extending our understandings about the high energy activity around the AGN jet in the Galaxy.

5

Prospects and problems

Existence of high energy, non-thermal electrons already came to our notice about a half century ago as a result of radio observation. The radiation from energetic protons emerges only in the 7-ray energy band, and those objects which are relevant to the origin of cosmic rays are naturally the ' archetype' 7-ray source. As one of such objects, SNR is generally accepted as the most likely object of accelerating cosmic rays up to ~ 1015 eV, however not beyond this energy. In addition to SNR, unknown site of the origin of cosmic rays must exist and remains to be seeked for. A new class of VHE 7-ray sources is to be found.

39 5.1

Origin of cosmic rays

Several unidentified EGRET 7-ray sources are accompanied by SNR, and such examples in the northern sky are studied by the Whipple, CAT and HEGRA groups to detect VHE 7-rays extrapolated from the GeV flux with a power index of ~2 as predicted from the shock acceleration model of cosmic ray pro­ tons. The efforts are so far negative in contrast with the result from SN 1006. The both SNRs of CANGAROO detection appear dissimilar to those SNRs in the unidentified EGRET sources; considerably dark in radio and GeV 7-rays but bright in non-thermal X-rays. The fact can be consistent with VHE 7-rays emitted from electrons and suggest that the component from proton progenitor is weak below the detection sensitivity. The shock acceleration of electrons in the shell of SNR is proved from the SN 1006 result. However, the path that will take us to understand the origin of cosmic rays is somewhat twisted from the standard expectation that detection of VHE 7-rays will directly provide evidence for acceleration of cosmic ray protons in SNR. The acceleration of cosmic ray protons and the subsequent 7ray emission depend on the complexity of various types of SNRs that may stem from environmental conditions. Emission of 7-rays may be modified from the earlier considerations (Drury et al. 1994; Naito and Takahara 1994) to allow for a variety of the characteristics which individual SNRs appear to show, as well as to refine the models of acceleration and radiation mechanism. It is argued that the confinement life time of cosmic rays in a cluster of galaxies is much longer than ~ 107 years, the time in our Galaxy, and even as long as the Hubble time (Volk et al. 1996). If so, cosmic ray interaction in a cluster of galaxies may give a VHE 7-ray flux which is detectable by using the system of Cerenkov telescopes. The star burst galaxies are likely to contain intense cosmic rays, and are also putative sources. The 7-ray observation of other galaxies would give new insights about the activities of high energy processes like supernova explosions during earlier periods and then their effects to the evolution of galaxies. 5.2

New population of sources?

The activities of AGN and pulsars which are the identified 7-ray sources in GeV and TeV energy region, are due to the compact object of black hole and neutron star. However, not all types of the compact objects are found to be 7-ray sources, as noted in Table 2. The emission from the compact objects appear violently time-variable sources, except for single pulsar, as observed in the other wave bands for instance X-rays. There is in any wave bands no evidence observed on single black holes (not in a binary system) and neutron

40 Table 2: Compact objects and 7-ray sources.

type

neutron star

Galactic candidates of black hole of M ~ 1QMQ

Black holes in the center of galaxies

single

^

no evidence

\J

in any wave band binary

^ (? ;Cen X-3)

(X-ray source)

(colliding galaxies !?)

The mark ^ indicates detection by 7-rays stars after its death as radio pulsar. The maximum energy of acceleration of electrons is likely determined by the radiation loss (Ghisellini et al. 1998), and the average energy density, for instance, of radiation field, in the vicinity of compact accreting objects de­ pends on the mass M of compact star (the spatial size and the brightness or the Eddington luminosity generally changes in scale to the mass M of com­ pact star, so that the energy density is generally proportional to the inverse of mass, oc M~l) as well as the amount of energy input supplied to the object. In the case of pulsars, accretion is the energy source for a pulsar in binary system. The size of radiation region is presumably no larger than the binary orbit, and less than 1 pc of the nebula size of single pulsars, giving the spatial density of radiation energy as higher in the case of binary system. For pro­ tons as the progenitor of 7-rays, the region where radiation and acceleration take place can be considerably separated from each other, because protons' radiation life time against ambient electromagnetic field is much longer than electrons' and the matter density required for 7r° production may be low in the acceleration region. The condition on the optical thickness against pair production is different from the electron. Latent sources of VHE 7-rays in a binary system, for instance Cen X3, have environmental conditions different from single pulsar and AGN, and thus likely with an emission mechanism dissimilar to the known 7-ray sources. However, investigation on such objects so far done is far from suffiient and complete. The unidentified EGRET sources of violent time-variability suggest a new population of 7-ray sources (Tavani et al. 1998). The time-variabile sources may include Galactic micro quasar which accompanies a relativistic jet to enhance high energy radiation like the case of AGN and 7-ray bursts.

41

Though these objects will be surveyed by using Cerenkov imaging telescope of the next generation, the field of view typically of 3° - 5° across is rather suited to pin-pointed deep study of selected objects. Cooperative informa­ tion from other bands are necessary for properly choosing observation targets such as violently time-variable sources as well as for simultaneous multi-bands observation and interpretation. Exotic interaction of particles generally has an energy threshold and can appear only at higher energies. A line emision from annihilating dark matters might exist in 100 GeV ~ 1 TeV region, and may be detectable with the energy resolution AE7/Ey ~20% which is expected for the system of mutiple telescopes. Increasing deviation of propagation speed of photons from the light velocity c might happen with increasing photon energies (Amelino-Camelia et al. 1998; Coleman and Glashow 1997), and the effect would modify the travelling distance of 7-rays (Kifune 1999). Extragalactic diffuse 7-rays have been measured up to about 50 GeV by EGRET. Extrapolation of the spectrum to higher energies and observation of VHE 7-ray emission from the halos of galaxies would provide a test for these exotic phnomena. 6

Conclusion

Observation of VHE 7-rays will extend into the energy region of ~ 100 GeV by using a system of ground-based multiple telescopes. A greater number of AGN at larger distances will be detected. Although investigation will be still limited to small values of redshifts in the order of ~ 0.1, the conversion effect itself of 7-rays to electron and positron pair provides us with the means of estimating the infrared intensity in extragalactic space. Efforts for understanding parti­ cle acceleration in AGN jet, pulsar wind and shock process will be continued by investigating more samples of AGN, pulsars and SNRs with more samples. Various types of objects are characterized by different values of, for instances, mass or luminosity which distrubutes in a wide range of the parameters. In unexploited values of these, a new population of 7-ray sources might emerge with different schemes of particle acceleration. The objects where proton ac­ celeration takes place still remain to be looked for. The study is linked to new populations of 7-ray sources, and extended to the case of extragalactic cosmic rays. As located in the highest energy band of electromagnetic radiation, tests for speculated exotic phenomena are a unique role to be done by VHE 7-ray astronomy. Cosmic rays are known to be the highest energy phenomenon in the present Universe, and VHE 7-rays are to extend the survey of such activity beyond our Galaxy.

42

References 1. G. Amelino-Camelia, J. Ellis, N.E. Mavromatos, D.V. Nanopoulos, and S. Sarkar, Nature 393, 763 (1998). 2. P.M. Chadwick et al, APJ 503, 391 (1998) 3. S. Coleman and S.L. Glashow, Phys L B405, 249 (1997) 4. L.O'C. Drury, F.A. Aharonian and H.J. Volk, AA 267, 959 (1994) 5. R.C. Hartrnan et al, APJS 123, 79 (1999) 6. G. Ghisellini et al, MNRAS 301, 451 (1998) 7. T. Kifune, APJ 518, L21 (1990) 8. T. Kifune, Adv,Space Res. 25, 641 (2000) 9. M. Mori, Proc. of TeV gamma-ray Workshop (Snowbird, Utah; August 1999) , (1999) 10. H. Muraishi et al, AA i, n (p)ress 2000 11. T. Naito and F. Takahara, J. Phys G 20, 477 (1994) 12. M. Tavani et al, APJ 497, L89 (1998) 13. W.T. Vestrand, P. Sreekumar and M. Mori, APJ 483, L49 (1997) 14. H.J. Volk, F.A. Aharonian and D. Breitschwerdt, Space Science Review 75, 279 (1996) 15. T.C. Weekes, F.A. Aharonian, D.J. Fegan and T. Kifune, Proc. of 4th CGRO Symposium (The American Institute of Physics, edited by C D . Dermer, M.S. Strickman and I.D. Kurfess; AIP Conference Proceedings 410) , 361 (1999)

Evolution and Explosion of Massive Pop III Stars and their Nucleosynthesis Hideyuki Umeda, Marii Shirouzu and Ken'ichi Nomoto Research

Center for the Early Universe and Department of Astronomy, of Tokyo, Hongo, Bunkyo-ku, 113-0033, Japan

University

We calculate stability, evolution and explosion of massive Pop III stars. We find the upper limit mass of the pulsationally stable Pop III ZAMS stars is about 13OM0, and the mass loss rate of unstable stars may be low. The nucleosynthesis results are compared with abundances of metal-poor halo stars to constrain the IMF of Pop III stars. The interesting trends of the observed ratios [Zn, Co, Mn, Cr/Fe] of the very metal-poor stars can be related to the mass ratio between the complete Si burning region and the incomplete Si burning region. We find that yields of Type II supernovae and Fe core collapse hypernovae can be consistent with the very low metal star abundances, if significant amount of complete Si-burning products are mixed out in the ejecta.

1

Introduction

Massive Pop III stars are important since they make the first metal enrichment of the universe, which affects significantly the following chemical and dynamical evolution of galaxies. There are several suggestions that the IMF of Pop III is different from Pop I and II, and the typical mass range of the first stars is likely to be 100 - 200M© (e.g., Nakamura & Umemura 1999, Ferrara 2000, Coppi 2000). If this is the case, the abundance pattern of the early galaxy should be similar to the yield of these stars. Recent observations have shown that the abundance pattern of halo stars are different for [Fe/H] < —2. Specifically, the Co/Fe and Zn/Fe ratios in­ crease significantly for smaller metallicity while Mn/Fe and Cr/Fe decreases (McWilliam et al. 1995, Ryan et al. 1996, Primas et al. 2000). These inter­ esting trends could be related to the IMF of Pop III stars. We have discussed that decreasing trend of Mn, Cr and increasing trend of Co can be explained simultaneously if the mass cut that divides the ejecta and the compact remnant tends to be deeper for massive core collapse supernovae (SNe) (Nakamura et al. 1999). This is because Mn and Cr are produced mainly in the incomplete Si burning region, while Co is produced in the deeper complete Si burning region. The mass cut is typically located somewhere close to the border of complete and incomplete Si burning. Therefore, if the mass cut is deeper, the abundance ratio of Co/Mn increases. Zn is also produced chiefly in the complete Si burning region, and our theory predicts increaseing trend of [Zn/Fe] below [Fe/H]~ —2.5, which matches with recent observations

43

44 Table 1: The results of the stability analysis for Pop III and Pop I stars. O a n d x represent that the star is stable and unstable, respectively. The e-folding time for the fundamental mode is shown after x in units of 10 4 yr. mass (M@) Pop III Pop I

80

100

o o

O

x (7.02)

120

o

x (2.35)

150 x (9.03) x (1.43)

180 x (4.83) x (1.21)

300 x (2.15) x (1.71)

(Primas et al. 2000, Ryan 2000). We study the nucleosynthesis pattern of massive Pop III stars with various masses to compare with abundances of metal-poor halo stars, and we try to constrain the typical mass range of Pop III stars that enriched the early galaxy. 2

Stability and Mass Loss of Massive Pop III stars

To determine the upper limit mass of the Zero Age Main Sequence (ZAMS), we analize a linear non-adiabatic stability of massive (SOMQ - 3OOM0) Pop III stars using a radial pulsation code (Shibahashi 2000). Stellar structures are calculated by the classical fitting method. We adopt e~ scattering and Kramers formula for the opacity, and the main nuclear reactions are pp-chain, hot-CNO cycle, and 3a reaction. Because the CNO elements are absent during the early stage of their evolution, the CNO cycle does not operate and the star contracts until temperature rises sufficiently high for the 3a reaction to produce 1 2 C. We calculate that these stars have X c n 0 ~ 1.6 — 4.0 x 10~ 10 , and the central temperature Tc ~ 1.4 x 108K on thier ZAMS. We also examine the models of Pop I stars for comparison. Table 1 shows the results for our analysis. The critical mass of ZAMS Pop III star is 128M 0 while that of Pop I star is 94M 0 . This difference comes from very compact structures (with high Tc) of Pop III stars. Stars more massive than the critical mass will undergo pulsation and mass loss. We note that the e-folding time of instability is much longer for Pop III stars than Pop I stars with the same mass, and thus the mass loss rate is much lower. These results are consistent with Ibrahim et al (1981) and BrafFe et al (2000). We thus assume that the mass loss is negligible in Pop III stars in the following calculations. 3

Stellar Evolution and Nucleosynthesis

We calculate the evolution of massive Pop III stars from ZAMS to SN explo­ sion. Stellar evolution is calculated with a Henyey type stellar evolution code

45

(Umeda et al. 2000 ; see also Nomoto & Hashimoto 1998) and SN explosion is simulated with a PPM code. The nucleosynthesis during the explosion is calculated by postprocessing (Nakamura et al. 2000). The stellar evolution and the mechanism of SNe explosion depend on the stellar mass. In the following, we classify these stars into four types, and describe the nucleosynthesis results for each one. 3.1

M ~ 10 - 25M 0 (Type II SNe)

Stars in the mass range M ~ 10 —25M 0 explode as ordinary Type II SNe (SNe II) with explosion energies jBexp ~ 10 51 erg. The abundance distributions after explosion for 13 and 25 MQ models are shown in Figures 1 and 2, respectively. In the ejecta, the mass fraction of complete Si burning products is larger if mass-cut is deeper (i.e., M c u t is smaller). In Figures 1 and 2 we show [Zn, Co, Mn/Fe] in the ejecta as a function of M c u t and the 56 Ni mass for the M c u t . We find that large [Zn, Co/Fe] and small [Mn/Fe] can be realized simulta­ neously with small M c u t . For example, in the 13M© model for M c u t ~ 1.55M 0 , [Zn/Fe] is large enough to explain the observed large ratio [Zn/Fe]~ 0.5. More massive stars also can realize large [Zn/Fe] if mass-cut is deeper. We should note that the ejected 56 Ni mass increases for deeper mass-cuts. Actually bright supernovae with larger 56 Ni mass ejection have been observed (Nomoto et al. 2000). However, the 56 Ni mass required to get [Zn/Fe]~ 0.5 appears to be too large in comparing with observations of [O/Fe]. However, as will be discussed in §4.2, the 56 Ni mass can be smaller without changing the [Zn/Fe] ratio if fall-back occurs after mixing. 3.2

M ~ 25 - 13OM0 (Ft Core Collapse Hypernovae)

Recently we had some evidences that at least some core collapse SNe explode with large explosion energy, which may be called "Hypernovae" (e.g., Nomoto et al. 2000). These SNe likely originate from relatively massive SNe (M > ~ 25M 0 ). In Figure 3 we show nucleosynthesis in the 25M 0 stars with the explosion energy 1052 erg to compare with the 10 51 erg model in Figure 2. We find that for larger explosion energy, boundaries of both complete and incomplete Si burning regions move outward in mass coordinate. We also find that for larger explosion energy incomplete Si burning region is thicker in mass. Therefore, hypernova explosions can eject a large amount of products from complete Si burning region only if more 56 Ni are ejected than ordinary SNe II. We also show [Zn, Co, Mn/Fe] and 56 Ni mass in the ejecta as a function of M c u t in Figure 3. Figure 3 indicates that if the ejected 56 Ni mass is fixed to

46 1

1

'

'

'

I

i

1

0.5(im) to obtain data with reasonable errors. Isotopic studies of presolar grains provide us detailed information which cannot be obtained by any other method. In the following sections, several cases are discussed. 2

Mainstream SiC grains

Of various types of presolar grains, SiC has been extensively studied by several reasons. First, concentrations of SiC in meteorites are relatively high [20]. In the Murchison meteorite, the concentration of SiC is 6ppm, while that of graphite is about lppm [21]. Second, SiC grains have been observed in different types of meteorites, while presolar graphite has been observed only a handful of meteorites [20]. In addition, they have relatively high trace element concentrations, making it possible to obtain isotopic ratios of those elements [22].

2. /

Light element isotopes

Figure 1 shows C and N isotopic ratios of SiC grains analyzed by SIMS [23]. According to their C, N, and Si isotopic ratios, several populations can be distinguished. The mainstream grains comprise about 94% of SiC grains [24]. These, along with Y and Z grains, are believed to have formed in AGB stars [25, 26]. X grains are inferred to have a SN origin [10, 27]. The origin of A and B grains are not clear. It should be noted that the distribution of the grains in Fig. 1 does not

57

reflect their natural proportions, since rare types of grains were search by ion imaging techniques and are overrepresented.

Fig. 1 Carbon and N isotopic ratios measured in individual SiC grains from the Murchison meteorite. Five populations can be distinguished on the basis of the C, N, and Si isotopic ratios. Also indicated are the theoretically expected ranges resulting H and He burning and subsequent dredge-up and from extra mixing in low-to intermediate carbon stars.

It has been widely accepted that mainstream grains formed in thermally pulsing AGB stars [23, 28]. Bulk analyses have shown that isotopic ratios of trace elements, Kr [14, 15], Xe [14, 15], Sr [16], Ba [17-19], Nd, and Sm [18, 29] have s-process signatures, while AGB stars are considered to be a main s-process operating site. I2 13 C/ C ratios of mainstream SiC grains range from 30 to 100 and are in agreement with the observational values of C-rich AGB stars. Finally, 11.2(im feature which is characteristic of SiC has been observed in C-rich AGB stars [30]. Carbon and N isotopic ratios of most mainstream grains agree with models of AGB stars, if cool bottom processing (CBP) [31] or extra mixing [32, 33] is taken into account. However, low l4N/'5N ratios ( 1 for the mix to form carbonaceous grains. Travaglio et al. [47] have performed mixing calculations in order to quantitatively examine the mixing and to reproduce isotopic data of low-density graphite grains and other SN grains by using a Type II supernova model by Woosley and Weaver [51]. We will briefly summarize their results and discussions here.

Fig. 5 Travaglio et al. [47] divided a 15M supernova by Woosley and Weaver [51] into seven different zones. The zones Ml, M2, and M3 have C>6, zones M4 and M5 are dominated by 160 and zone M6 is dominated by 28Si, and zone M7 is dominated by 56Ni.

They divided a supernova into 7 zones, Ml, M2, M3 being He-rich, M4 and M5 being O-rich, M5 and M6 being Si-rich, and M7 being Ni-rich (Fig. 5). They mixed those zones so that the resulting mix should have the condition C/0>1 and

63 12

C/I3C ratios would be in the range of 5 to 104. The contribution from zone M4 is fixed and that of M5 and M6 are changed. Overall trend and the ranges of the isotopic data can be reproduced by the mixing (see Fig. 4). However, several isotopic features which cannot be explained by the model and the major problems are summarized below. 3.1

,5

N deficiency

If all the 14N in M1 (He/N) zone is incorporated in the mix, the N-isotopic ratios become much higher than those observed in the grains. Thus, they arbitrarily reduced the amount of 14N in the Ml zone by a factor of 20 (Fig. 4a). Their result strongly suggests that either 15N in the He/C zone is deficient in the model or most l4 N in the He/N zone does not go into the grains. The former is quite possible. Langer [56] examined rotational mixing in massive stars and found that by mixing proton into He-rich region, 15N production can be substantially higher than previously predicted. In addition, 15N production during explosive nucleosynthesis strongly depends on the strength of the shock. Thus, the l5N deficiency may be resolved once these effects are taken into account. 3.2

29

Si deficiency

As shown in Fig. 4d, the mix is short of 29Si to reproduce the grain data, most grains being plotted above the lines. It suggests that 29Si is underproduced in the model of Type II. This is strengthened by the fact that Timmes and Clayton [36] failed to explain the cosmic ratio of MSi/wSi (1.7) by using productions of different stellar sources including Type II supernovae. The deficiency of 29Si relative to MSi is also observed in another SN model by Thielemann et al. [57] and it may be a problem of fundamental parameters, such as reaction rates, in the models. Travaglio et al. [47] pointed out that 26Mg(a,n)29Si reaction rate can be higher by a factor of 2, possibly accounting for the 29Si deficiency. It remains to be seen whether this assumption can be justified. 4

Summary

As shown in the previous sections, isotopic analyses of presolar grains can provide the detailed information on nucleosynthesis inside stars and conditions of supernova explosion. Mainstream grains most likely formed in thermal-pulsing AGB stars. Overall isotopic ratios of the grains can be explained in the framework of the models of AGB stars. However, low l4N/lsN ratios of a few grains are difficult to explain by the models. The Si isotopic ratios reflect both initial isotopic compositions of the

64

parent stars and nucleosynthesis in the He-shell. Isotopic ratios of heavy elements of the grains are dominated by a component from the He-shell. s-Process isotopic ratios in the He-shell inferred from the grain data agree with what is predicted in low-mass AGG stars of solar metallicity. Silicon carbide type X, low-density graphite, and silicon nitride have a distinct isotopic feature which indicates that they formed in supernova ejecta. In order to quantitatively explain isotopic ratios of the grains, Travaglio et al. [47] calculated mixing between different zones of Type II supernova models by Woosley and Weaver [51]. The mixing model can explain general isotopic features if material from the Si-rich zones is assumed to penetrate the O-rich zones and is mixed with that of the C-rich zones. The major problems are that l5N is underproduced in the model and that 29Si/28Si ratios of the mixing are consistently lower than those in the grains. These problems can be overcome by multi-dimensional models and/or adjustment of cross sections. 5

Acknowledgements

I thank the organizing committee for inviting me to attend the OMEG2000 conference. I am grateful to Ernst Zinner for his continuous encouragement throughout the study of presolar grains. This work has been supported by NASA grant NAG5-8336. References 1. Lewis R. S., Tang M., Wacker J. F., Anders E. and Steel E., Interstellar diamonds in meteorites, Nature 326, (1987) pp. 160-162. 2. Black D. C. and Pepin R. O., Trapped neon in meteorites. II., Earth Planet. Sci. Lett. 6, (1969) pp. 395-405. 3. Black D. C , On the origins of trapped helium, neon and argon isotopic variations in meteorites II. Carbonaceous meteorites, Geochim. Cosmochim. Acta 36, (1972) pp. 377-394. 4. Anders E., Circumstellar material in meteorites: noble gases, carbon and nitrogen, In Meteorites and the Early Solar System, ed. by J. F. Kerridge and M. S. Matthews (University of Arizona Press, 1988) pp. 927-955. 5. Bernatowicz T., Fraundorf G., Tang M., Anders E., Wopenka B., Zinner E. and Fraundorf P., Evidence for interstellar SiC in the Murray carbonaceous meteorite, Nature 330, (1987) pp. 728-730. 6. Tang M. and Anders E., Isotopic anomalies of Ne, Xe, and C in meteorites. II. Interstellar diamond and SiC: carriers of exotic noble gases, Geochim. Cosmochim. Acta 52, (1988) pp. 1235-1244.

65

7. Amari S., Anders E., Virag A. and Zinner E., Interstellar graphite in meteorites, Nature 345, (1990) pp. 238-240. 8. Lewis R. S. and Srinivasan B., A search for noble-gas evidence for presolar oxide grains, Lunar Planet. Sci. XXIV, (1993) pp. 873-874. 9. Nittler L. R., Alexander C. M. O'D., Gao X., Walker R. M. and Zinner E. K., Interstellar oxide grains from the Tieschitz ordinary chondrite, Nature 370, (1994) pp. 443-446. 10. Nittler L. R., Hoppe P., Alexander C. M. O'D., Amari S., Eberhardt P., Gao X., Lewis R. S., Strebel R., Walker R. M. and Zinner E., Silicon nitride from supernovae, Astrophys. J. 453, (1995) pp. L25-L28. 11. Nittler L. R., Alexander C. M. O'D., Gao X., Walker R. M. and Zinner E., Stellar sapphires: The properties and origins of presolar A120;, in meteorites, Astrophys. J. 483, (1997) pp. 475-495. 12. Bernatowicz T. J., Cowsik R., Gibbons P. C., Lodders K., Fegley B., Jr., Amari S. and Lewis R. S., Constraints on stellar grain formation from presolar graphite in the Murchison meteorite, Astrophys. J. 472, (1996) pp. 760-782. 13. Bernatowicz T. J., Amari S., Zinner E. and Lewis R. S., Interstellar grains within interstellar grains., Astrophys. J. 373, (1991) pp. L73-L76. 14. Lewis R. S., Amari S. and Anders E., Meteoritic silicon carbide: pristine material from carbon stars, Nature 348, (1990) pp. 293-298. 15. Lewis R. S., Amari S. and Anders E., Interstellar grains in meteorites: II. SiC and its noble gases, Geochim. Cosmochim. Acta 58, (1994) pp. 471494. 16. Podosek F. A., Prombo C. A., Amari S. and Lewis R. S., s-Process Sr isotopic compositions in presolar SiC from the Murchsion meteorite, Astrophys. J. (2000) in press. 17. Ott U. and Begemann F., Discovery of s-process barium in the Murchison meteorite, Astrophys. J. 353, (1990) pp. L57-L60. 18. Zinner E., Amari S. and Lewis R. S., s-Process Ba, Nd, and Sm in presolar SiC from the Murchison meteorite, Astrophys. J. 382, (1991) pp. L47-L50. 19. Prombo C. A., Podosek F. A., Amari S. and Lewis R. S., s-Process Ba isotopic compositions in presolar SiC from the Murchison meteorite, Astrophys. J. 410, (1993) pp. 393-399. 20. Huss G. R. and Lewis R. S., Presolar diamond, SiC, and graphite in primitive chondrites: Abundances as a function of meteorite class and petrologic type, Geochim.. Cosmochim. Acta 59, (1995) pp. 115-160. 21. Amari S., Lewis R. S. and Anders E., Interstellar grains in meteorites: I. Isolation of SiC, graphite, and diamond; size distributions of SiC and graphite, Geochim. Cosmochim. Acta 58, (1994) pp. 459-470.

66

22. Amari S., Hoppe P., Zinner E. and Lewis R. S., Trace-element concentrations in single circumstellar silicon carbide grains from the Murchison meteorite, Meteoritics 30, (1995) pp. 679-693. 23. Hoppe P., Amari S., Zinner E., Ireland T. and Lewis R. S., Carbon, nitrogen, magnesium, silicon and titanium isotopic compositions of single interstellar silicon carbide grains from the Murchison carbonaceous chondrite., Astrophys. J. 430, (1994) pp. 870-890. 24. Hoppe P. and Ott U., Mainstream silicon carbide grains from meteorites, In Astrophysical Implications of the Laboratory Study of Presolar Materials, ed. by T. J. Bernatowicz and E. Zinner (AIP, New York, 1997) pp. 27-58. 25. Hoppe P., Annen P., Strebel R., Eberhardt P., Amari S. and Lewis R. S., Circumstellar SiC grains of Type Z: Evidence for extensive He shell dredge-up in low-metallicity low-mass AGB stars, Lunar Planet. Sci. XXVIII, (1997) pp. 599-600. 26. Amari S., Nittler L. R., Zinner E., Gallino R., Lugaro M. and Lewis R. S., Presolar SiC grains of Type Y: Origin from low-metallicity AGB stars, Astrophys. J. (2000) submitted. 27. Amari S., Hoppe P., Zinner E. and Lewis R. S., Interstellar SiC with unusual isotopic compositions: Grains from a supernova?, Astrophys. J. 394, (1992) pp. L43-L46. 28. Huss G. R., Hutcheon I. D. and Wasserburg G. J., Isotopic systematics of presolar silicon carbide from the Orgueil (CI) carbonaceous chondrite: Implications for solar system formation and stellar nucleosynthesis., Geochim. Cosmochim. Ada 61, (1997) pp. 5117-5148. 29. Richter S., Ott U. and Begemann F., S-process isotope abundance anomalies in meteoritic silicon carbide: new data, In Nuclei in the Cosmos, ed. by F. Kappeler and K. Wisshak (Institute of Physics Publishing, Bristol and Philadelphia, 1993) pp. 127-132. 30. Little-Marenin I. R., Carbon stars with silicate dust in their circumstellar shells, Astrophys. J. 307, (1986) pp. L15-L19. 31. Wasserburg G. J., Boothroyd A. I. and Sackmann I.-J., Deep circulation in red giant stars: A solution to the carbon and oxygen isotope puzzles?, Astrophys. J. 447, (1995) pp. L37-L40. 32. Charbonnel C , Clues for non-standard mixing on the red giant branch from 12 13 C/ C and l2C/14N ratios in evolved stars, Astron. Astrophys. 282, (1994) pp. 811-820. 33. Charbonnel C , A consistent explanation for 12C/13C, 7Li, and 3He anomalies in red giant stars, Astrophys. J. 453, (1995) pp. L41-L44. 34. Forestini M., Paulus G. and Arnould M., On the production of 26A1 in AGB stars, Astron. Astrophys. 252, (1991) pp. 597-604. 35. Gallino R., pers. comm. (2000)

67

36. Timmes F. X. and Clayton D. D., Galactic evolution of silicon isotopes: Application to presolar SiC grains from meteorites, Astrophys. J. All, (1996) pp. 723-741. 37. Clayton D. D. and Timmes F. X., Placing the Sun in galactic chemical evolution: Mainstream SiC particles, Astrophys. J. 483, (1997) pp. 220227. 38. Lugaro M., Zinner E., Gallino R. and Amari S., Si isotopic ratios in mainstream presolar SiC grains revisited, Astrophys. J. 527, (1999) pp. 369-394. 39. Nicolussi G. K., Davis A. M , Pellin M. J., Lewis R. S., Clayton R. N. and Amari S., s-Process zirconium in presolar silicon carbide grains, Science 111, (1997) pp. 1281-1283. 40. Nicolussi G. K., Pellin M. J., Lewis R. S., Davis A. M., Amari S. and Clayton R. N., Molybdenum isotopic composition of individual presolar silicon carbide grains from the Murchison meteorite, Geochim. Cosmochim. Acta 62, (1998) pp. 1093-1104. 41. Gallino R., Raiteri C. M. and Busso M., Carbon stars and isotopic Ba anomalies in meteoritic SiC grains., Astrophys. J. 410, (1993) pp. 400-411. 42. Beer H., Voss F. and Winters R. R., On the calculation of Maxwellianaveraged capture cross sections, Astrophys. J. Suppl. 80, (1992) pp. 403424. 43. Voss F„ Wisshak K., Guber K., Kappeler F. and Reffo G., Stellar neutron capture cross sections of the Ba isotopes, Phys. Rev. C 50, (1994) pp. 25822601. 44. Koehler P. E., Spencer R. R., Winters R. R., Guber K. H., Harvey J. A., Hill N. W. and Smith M. S., Resonance neutron capture and transmission measurements and the stellar neutron capture cross sections of 134Ba and 136Ba, Phys. Rev. C 54, (1996) pp. 1463-1477. 45. Koehler P. E., Spencer R. R., Guber K. H., Winters R. R., Raman S., Harvey J. A., Hill N. W., Blackmon J. C , Bardayan D. W., Larson D. C , Lewis T. A., Pierce D. E. and Smith M. S., High resolution neutron capture and transmission measurement on 137Ba and their impact on the interpretation of meteoritic barium anomalies, Phys. Rev. C 57, (1998) pp. R1558-R1561. 46. Amari S., Zinner E. and Lewis R. S., Large 180 excesses in interstellar graphite grains from the Murchison meteorite indicate a massive star origin, Astrophys. J. 447, (1995) pp. L147-L150. 47. Travaglio C , Gallino R„ Amari S., Zinner E., Woosley S. and Lewis R. S., Low-density graphite grains and mixing in type II supernovae, Astrophys. J. 510, (1999) pp. 325-354. 48. Amari S. and Zinner E., Supernova grains from meteorites, In Astrophysical Implications of the Laboratory Study of Presolar Materials, ed. by T. J. Bernatowicz and E. Zinner (AIP, New York, 1997) pp. 287-305.

68

49. Nittler L. R., Amari S., Zinner E., Woosley S. E. and Lewis R. S., Extinct 44 Ti in presolar graphite and SiC: Proof of a supernova origin, Astrophys. J. 462, (1996) pp. L31-L34. 50. Hoppe P., Strebel R., Eberhardt P., Amari S. and Lewis R. S., Type II supernova matter in a silicon carbide grain from the Murchison meteorite, Science 272, (1996) pp. 1314-1316. 51. Woosley S. E. and Weaver T. A., The evolution and explosion of massive stars, II. Explosive hydrodynamics and nucleosynthesis, Astrophys. J. Suppl. 101, (1995) pp. 181-235. 52. Woosley S. E. and Weaver T. A., Sub-chandrasekhar mass models for type la supernovae, Astrophys. J. 423, (1994) pp. 371-379. 53. Clayton D. D., Arnett W. D., Kane J. and Meyer B. S., Type X silicon carbide presolar grains: Type la supernova condensates?, Astrophys. J. 486, (1997) pp. 824-834. 54. Clayton D. D., Liu W. and Dalgarno A., Condensation of carbon in radioactive supernova gas, Science 283, (1999) pp. 1290-1292. 55. Meyer B. S., Weaver T. A. and Woosley S. E., Isotope source table for a 25 Msun supernova, Meteoritics 30, (1995) pp. 325-334. 56. Langer N., Heger A., Woosley S. E. and Herwig F., Nucleosynthesis in rotating stars, In Nuclei in the Cosmos V, ed. by N. Prantzos (Editions Frontieres, Paris, 1998) pp. 129-135. 57. Thielemann F.-K., Nomoto K. and Hashimoto M.-A., Core-collapse supernovae and their ejecta, Astrophys. J. 460, (1996) pp. 408-436.

NUCLEOSYNTHESIS OF M G A N D AL IN GLOBULAR CLUSTER R E D G I A N T STARS C. I L I A D I S Carolina, Chapel Hill, NC 27599, USA and Universities Nuclear Laboratory, Durham, NC 27708, USA E-mail: [email protected]

The University Triangle

of North

The proton-capture reaction on 2 4 M g has been investigated in the bombarding energy range of E p =0.2-1.7 MeV. Properties of low-energy resonances have been measured. From the experimental results, accurate proton partial widths, 7-ray partial widths and total widths have been deduced. The present experimental information establishes the 2 4 M g + p reaction rates over the temperature range T=0.02-2.0 GK with statistical uncertainties of less than 21%. Based on our results, we can rule out the recent suggestion that the total width of the E/j=223 keV resonance has a significant influence on the reaction rates. The astrophysical implications for hydrogen burning of 2 4 Mg at low stellar temperatures in globular cluster red giant stars are discussed.

1

Introduction

Standard stellar evolution theory predicts for low-mass stars in globular clus­ ters only minor changes in surface composition during the red giant branch phase. However, observations over the past two decades have shown a large variation in the abundances of C, N, O, Na, Mg and Al among globular cluster red giant stars l . These observations have been explained by hydrogen burning nucleosynthesis converting O to N, Ne to Na, and Mg to Al. However, the site in which the proton captures take place has not yet been identified. The observed surface abundance variations might result from internal nucleosyn­ thesis and deep, but yet unexplained, mixing in the red giants (evolutionary scenario), or could be caused by the nucleosynthesis in stars of a previous gen­ eration that contaminated the proto-stellar cluster gas (primordial scenario). For the globular cluster M13 there is increasing observational evidence in favour of the evolutionary scenario 1. For example, the observed abundance sum Mg+Al is approximately constant in M13 red giants that exhibit large variations in Al. The implication is that all stars had initially the same abun­ dance of Mg, but experienced different degrees of Al production and of deep mixing at the same evolutionary state. Theoretical model calculations 2 ' 3 that are based on the evolutionary scenario are in good qualitative agreement with most of the observed light-element abundance variations and correlations. The

69

70

calculations predict that 25 Mg and 26 Mg are converted to Al at temperatures of T=0.02-0.06 GK in the hydrogen-burning shell. However, recent observations of Mg isotopic abundances for a number of M13 red giants by Shetrone 4 have found, contrary to expectations, that the abundance of 24 Mg is anticorrelated with Al, and that the abundance sum 25 Mg+ 26 Mg is approximately constant for large variations in Al. This result is difficult to explain, since current esti­ mates of reaction rates for proton captures on Mg isotopes 5 ' 6 predict that the rates for 2 5 Mg+p and 2 6 Mg+p are much larger than for 2 4 Mg+p. However, it was recently pointed o u t 7 that the 2 4 Mg+p reaction rates might be much larger than previously assumed due to the uncertain contribution of the lowenergy wing of the ER=223 keV resonance. Accurate calculation of the wing contribution requires knowledge of the proton and 7-ray partial widths of this resonance. Here, we report on the measurement of the strength w-y and the branching ratio T 7 / r of the E#=223 keV resonance in 2 4 Mg+p. In addition, we have measured the mean lifetime r m of the E x =2485 keV state in 25A1 using the DSA method. The determination of the two unknown resonance parameters T p and T 7 from the three measured quantities T 7 / r , ivy and r m represents an important test of internal consistency. In the following we discuss the experi­ mental equipment, the experimental results and the astrophysical implications. For more details, the reader is referred to Ref. 8 . 2

Experimental Equipment

The experiments were carried out at the Triangle Universities Nuclear Lab­ oratory (TUNL). A 4 MV modified KN Van de Graaff accelerator provided proton beams up to 10 /uA on target at energies of E p =0.9-2.0 MeV. A 200 kV minitandem accelerator supplied proton beams in the energy range E p {

:

/

■ l i t • ' I I I

■i '

A

/

' V / ;i/ /xi \

i DC / /

N

/

i

i

1 0"'

'

\ / ' / '/ \ / ' '"' / y / .'/ /

■

1 0"

^ - C ^ ? ^

\

i

l

/

823 keV

■

l

i

l

l

I

I

I

'■ :

ft /,' 1 :419 keV/ ■

'

i

/

'

/

/

E > 823 keV :

. . i . .. .ii i

10

R

i

10"

T(GK)

Figure 1: (a) Total reaction rate (solid line) and individual contributions of resonances (dashed lines) and direct capture (dotted line) for the reaction 2 4 Mg(p,7) 2 5 Al; (b) ratio of the present recommended total reaction rate to previous results of Ref. 5 .

76

Acknowledgments This work was supported in part by the U.S. Department of Energy under Contract No. DE-FG02-97ER41041. References 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14.

R.R Kraft et al, Astron. J. 115, 1500 (1998). G.E. Langer et al, Pub. Astron. Soc. Pac. 109, 244 (1997). R.M. Cavallo et al, Astrophys. J. 492, 575 (1998). M.D. Shetrone, Astron. J. 112, 2639 (1996). G.R. Caughlan and W.A. Fowler, At. Data Nucl. Data Tables 40, 283 (1988). C. Angulo et al, Nucl. Phys. A, in print (1999). C.S. Zaidins and G.E. Langer, Pub. Astron. Soc. Pac. 109, 252 (1997). D.C. Powell et al, Nucl. Phys. A, in print (2000). D.C. Powell et al, Nucl. Phys. A 644, 263 (1998). Trautvetter, H.-R, Nucl. Phys. A 243, 37 (1975). Blaugrund, A.E., Nucl. Phys. 88, 501 (1966). M. Uhrmacher et al, Nucl. Instr. Meth. B 9, 234 (1985). A.M. Lane and R.G. Thomas, Rev. Mod. Phys. 30, 257 (1958). P.A. Denissenkov et al, Astron. Astrophys. 333, 926 (1998).

X-RAY M E A S U R E M E N T S OF METAL A B U N D A N C E S OF HOT GAS IN CLUSTERS OF GALAXIES YASUSHI FUKAZAWA Department

of Physics, E-mail:

University of Tokyo, 7-3-1 Hongo, Bunkyo-ku, 113-0033, Japan [email protected]

Tokyo

Utilizing the imaging spectroscopic capabilities of the Japanease X-ray satellite ASCA, we measured Si and Fe abundances of 40 nearby clusters of galaxies. The spatially averaged Fe abundances of the intracluster medium (ICM) are 0.2-0.3 solar, with only weak dependence on the temperature of the intracluster medium, hence on the cluster richness. In contrast, the spatially averaged Si abundance is observed to increase from 0.3 to 0.6-0.7 solar from the poorer to richer clusters. These results suggest that the supernovae of both type-la and type-II significantly contribute to the metal enrichment of the intracluster medium, with the relative contribution of type-II supernovae increasing towards richer clusters. Many clusters exhibit a central increment in the Fe abundance, which is more pronounced in lower temperature clusters; +(0.1-0.2) solar at fcX > 5 keV, while +(0.2—0.3) solar at 1.5 < kT < 4 keV. These central excess metals are thought to be ejected from cD galaxies. Several low temperature cD type clusters also show significant Si abundance increase by +(0.1—0.2) solar at the central region. Compared with the Si-rich abundances observed in outer regions of rich clusters, the Si to Fe abundance ratio of central excess metals tends to be near the solar ratio, implying that type la products from cD galaxies are dominant in the central excess metals.

1

Introduction

Clusters of galaxies are quite popular objects, and contain 50-1000 member galaxies. Various optical observations indicate that clusters of galaxies are gravitationally bound systems. X-ray observations (e.g. Jones and Forman 1984) revealed that clusters of galaxies are filled with a large amount of X-ray emitting hot gas with the temperature of 10 7 ~ 8 K, which is called intracluster medium (ICM). The ICM mass is often 2-5 times massive than the total stellar mass in clusters of galaxies, and thus the ICM is an important constituent in clusters of galaxies and moreover in the universe. The ICM is thought to be gravitationally bound from various evidences obtained by X-ray observations (e.g. Jones and Forman 1984). The ICM contains a considerable amount of heavy elements, which is processed in the stellar interior, then ejected into interstellar/intergalactic space, and now accumulated in the ICM (Rothenflug and Arnaud 1985; Edge and Stewart 1991; Hatsukade 1989; Arnaud et al. 1992; Tsuru 1992). Therefore, significant fractions of heavy elements in the universe exist in clusters of galaxies, and the investigation of metal abundances

77

78

of the ICM is important to study the chemical evolution and star forming history in the universe. X-ray study of metal abundances of a cluster is in many respects more straightforward than the optical study of metal abundances of a galaxy. First, atomic lines observed in an X-ray spectrum from the ICM are emitted via physically much simpler processes than optical emission/absorption lines from stellar atmosphere. Second, the ICM is optically so thin that the X-ray lines are not affected by complex radiative transfer. Finaly, metal enrichment of the ICM is approximately "one-way" process wherein the feedback from the ICM to the member galaxies is negligible, whereas metals recycle in a very complicated manner within a galaxy between stars and interstellar gas. In fig­ ure 1, optically thin thermal bremsstrahlung spectra predicted by the plasma model (Raymond and Smith 1977) are shown. The Japanese fourth X-ray Observational satellite, ASCA (Tanaka et al. 1994), has an unprecedented energy resolution, a wide energy band 0.5-10 keV, and a moderate angular resolution 1-3', which are all essential for determining the temperature and metal abundances of extended hot gas. In particular, ASCA can resolve var­ ious atomic emission lines expected in the X-ray spectra of clusters. Before ASCA, only Fe abundance has been measured, and it is not spatially resoloved one.

Figure 1. Calculated X-ray spectra from optically thin hot plasmas with various temper­ atures. The Raymond-Smith plasma emission code (Raymond and Smith 1977) is used, assuming a metal abundance of 0.3 solar.

Utilizing ASCA, we can obtain metal abundances other than Fe. In this paper, in order to investigate the origin of metals in the ICM, we measured metal abundances of the ICM not only for Fe but also for Si as a function of the ICM temperature, together with their masses for a large sample clusters. Throughout this paper, we set the Hubble constant to be 50/iso km s _ 1 M p c - 1 . The detailed results are described in Fukazawa (1997), and main results are published in Fukazawa et al. (1998, 2000).

79

2

Observations and Data Reduction

We selected our sample clusters from the ASCA archival data, through the following rough criteria: the object must be extended enough to resolve spatially; it must be bright enough to constrain metal abundances; and it must not exhibit outstanding morphological asymmetry. The last criterion is necessary to avoid peculiar objects. In addition, we take attention so that clusters with various ICM temperatures can be included. Through these criteria, we have selected 40 clusters. Most objects are located at redshifts of z < 0.062. All the objects are brighter than 1 x 10~12erg s _ 1 c m - 2 in 0.5-10 keV which is enough to constrain metal abundances. Their ICM temperatures are distributed almost continuously from 1 to 10 keV. ASCA carries four indentical X-ray telescopes (XRT; Serlemitsos et al. 1995; Tsusaka et al. 1995) and two type focal plane imaging detectors with wide energy band 0.5-10 keV ; the GIS (Gas Imaging Spectrometer; Ohashi et al. 1996; Makishima et al. 1996) and the SIS (Solid-state Imaging Spectrometer), two detectors for both the GIS and SIS are onboard. The GIS has a moderate energy resolution 8% at 6 keV, and large field of view with the diameter 40', that covers the whole region of most clusters of galaxies. The SIS has a good energy resolution 120eV at 6 keV, good sensitivity in lower energy band, and smaller field of view 11' x 11' - 22' x 22'. The typical exposure time and count rate of each object are 20-80 ksec and 0.1-5 counts s e c - 1 per detector, respectively, and thus 10000-100000 photons were accumulated that is enough to mesaure the line strength. The background spectra were constructed from several blank-sky data, such as NEP, Draco, and QSF-3 fields for a total accumulation time of about 100 ksec, integrated in the same region as that of on-source spectra under the same data selection criteria. 3 3.1

Results Spatially Averaged Metal Abundances

Many clusters show a bright cool component and an abundance increase in the center (e.g. Fabian et al. 1994), which should be excluded from the present study of the average ICM properties. We therefore accumulated the GIS and SIS spectra for each cluster in a ring-shape region. The region typically has an inner radius of 0.1ftjj"0 Mpc and an outer radius of 0.4/ijT0 Mpc from the cluster center, respectively. In figure 2, we show an exmaple of ASCA spectra. Not only Fe-K lines but also Si, S, and Fe-L lines are detected.

80

■VV^/#/#^^ Energy (keV)

Energy

(keV)

Figure 2. The spatially averaged GIS/SIS spectra of A2199 cluster. The SIS data points exceed those of the GIS in lower energies; vise versa in higher energies. The solid line repre­ sents the best-fit single-temperature Raymond-Smith model assuming the solar abundance ratios, determined jointly by the two instruments. The model has been convolved through the XRT+SIS or XRT+GIS response to be compared with the SIS or GIS data. The left panel shows the whole band spectrum, and the right panel shows the spectrum around Si-K lines and solid lines represents the plasma model with the Si and S abundance to be 0 solar.

We perform combined fit to the GIS/SIS spectra with variable-abundance single temperature (IT) Raymond-Smith model (R-S model), that is typically utilized in X-ray astronomy. Solar abundances are taken from the solar photospheric values by Anders and Grevesse (1989), with (Fe/H) 0 = 4.68 x 10~ 5 and (Si/H)© = 3.55 x 10~ 5 . Free parameters of the fit are the interstellar absorption (A^H), temperature (kT), normalization, and the abundances of O, Mg, Si, S, and Fe. We asuume Ne abundances to be the same as O; Ca and Ar as S; and Ni as Fe, respectively, all in terms of the solar units. The abundances of He, C, and N are fixed at the solar values. For most clusters, the fit is acceptable with the reduced chi-square value of 1.4 or less with a typical degree of freedom 200. Thus, the spectra of most clusters can be represented fairly well by the IT R-S model with variable abundance ratios. In this way, we have determined spatially averaged metallicities of all clusters, with a typical accuracy of 10% or better for Fe and 40% for Si. Other elemental abundances are poorly constrained. We plot the derived Fe and Si abundances in figure 3a and 3b respectively, as a function of the ICM temperature. In figure 3a, the obtained Fe abundances distribute in a range of 0.2-0.4 solar, in rough agreement with the previous non-imaging results (e.g. Hatsukade 1989; Tsuru 1992). The Fe abundances of clusters with kT < 3 keV rely upon Fe-L lines. In order to evaluate the reliability of the Fe-L line results, we analyzed clusters with kT > 1.7 keV ignoring either Fe-K or FeL line regions of the spectrum, and obtained consistent results within 10%. This consistency is also reported by Mushotzky et al. (1996) and Hwang et

81

al. (1997) for several clusters. The ensemble-averaged Fe abundance is nearly independent of the temperature, at least over a range of 1.5 to 7 keV. This feature is common to different plasma codes utilized in the spectral fittings. The Fe abundance seems to decrease below kT ~ 1.5 keV. This trend is not convincing since the results depend on the plasma codes due to uncertainties in the Fe-L line modeling (Fabian et al. 1994; Fukazawa et al. 1996). As shown in figure 3b, Si abundance is 0.6-0.7 solar in rich clusters (kT > 4 keV), which is consistent with the previous ASCA measurements (e.g. Mushotzky et al. 1996). The present results thus extend their inference to a much larger sample. Moreover, in contrast with Fe, the Si abundance appears to correlate positively with the ICM temperature; it increases from 0.3 solar to 0.6-0.7 solar as the temperature increases, and it may saturate for kT > 4 keV.

1

2

5 10 1 2 5 10 kT (keV) kT (keV) Figure 3. Spatially averaged Fe and Si abundances excluding central regions, plotted as a function of the ICM temperature, (a) The Fe abundances of individual clusters obtained using the R-S code, (b) The same as panel (a), but for the Si abundance.

(D

Figure 4. Si to Fe abundance ratios aver­ aged over clusters with similar tempera­ tures, plotted as a function of the ICM tem­ perature. The different symbols denote dif­ ferent plasma codes.

o

S

kT (keV)

82

Finally, in figure 4, we present the sample-averaged Si to Fe abundance ratios in solar units, as a function of the ICM temperature. Figure 4 compares the R-S fitting results with those employing plasma emission codes by Masai (1984) and by Mewe-Kaastra-Liedahl (Mekal: Mewe et al. 1985; Liedahl et al. 1995): R-S, Masai, and Mekal codes are known to differ in the Fe-L line treatment (Fabian et al. 1994; Fukazawa et al. 1996). This ratio depends on the plasma code by 10-30%; the Mekal code gives the weakest correlation. Nevertheless, we can clearly see that the ratio positively correlates with the ICM temperature for all plasma code, increasing from ~ 1 at kT ~ 1 keV to ~2 at kT ~ 4 keV. This is the first detection of the variation of the abundance ratio in the ICM. Hwang et al. (1999) also reported that poor clusters exhibit lower Si to Fe abundance ratio than rich clusters. 3.2

Metal Abundances at the Central Cluster Region

We investigated the central 2' X-ray spectra to constrain metal abundances at the cluster central region. Here, we focus on clusters which contain a central dominant elliptical galaxy, since we are interested in the metal ejection from individual galaxies. It is found by ASCA that the averaged ICM temperature often decreases toward the cluster center (e.g. Fabian et al. 1994; Fukazawa et al. 2000). We thus try to fit the central spectra with two temperature R-S model (2T R-S model). In the 2T fitting, we fixed the temperature of the hot component to the spatially averaged cluster temperature. Metal abundances were assumed to be the same between the two components, since the data quality is usually not adequate to determine the abundances and normalization of the cool component independently. We plot the central Fe and Si abundances of clusters in figure 5 against the spatially averaged temperature. Many clusters, even hot ones, have higher Fe abundances than the spatially averaged values of 0.3 solar, and a clear negative correlation with the averaged ICM temperature is seen. This is in­ dependent of the plasma model. The Si abundance at the center also exhibits a noticeable difference from the outer-region values, instead of showing the positive correlation with the ICM temperature like in the outer region, the central Si abundance stays rather constant and high, 0.6-1.0 solar, except in the coolest objects. This is because cooler clusters exhibit stronger Si abun­ dance increases (by 0.2-0.3 solar) at the center, that is also seen by comparing figure 4 with figure 6 that plots the ensemble-avereged Si to Fe abundance ra­ tio. Finoguenov and Ponman (1999) also found the radial increase of Si to Fe abundance ratios of the ICM.

83 Fe Abundance in the center

Si Abundance in the center 1 1

I

I

t o


60) and the isotopic composition of ultraheavy particles (Z>30)in galactic cosmic rays as well as solar and interplanetary particles. The observation of nuclear composition covers a wide range of scientific themes including studies of nucleosynthesis of cosmic ray sources, chemical evolution of galactic material, the characteristic time of cosmic rays, heating and acceleration mechanism of cosmic ray particles. Observation of solar particle events also make clear the physical process of transient solar events emitting wide range of radio, X-ray/gamma-ray, plasma and energetic particle radiation, and particle acceleration mechanism driven by CME.

1

Scientific Objectives

The Cosmic Ray Isotope Observation for Nuclei Astrophysics (CORONA) program aims to measure nuclear composition of energetic particles observed by large scaled spacecrafts including Space Station, covering comprehensive studies of solar, interplanetary, anomalous cosmic ray (ACR) and galactic cosmic ray (GCR) particles [1,2]. Scientific objectives covering a wide range of basic problems are summarized as : A. Nuclear Composition, the Origin and Evolution of Galactic Matter : - To search for evidence of special sources of CRs, such as supernovae and WR-stars. - To study nucleosynthesis and galactic evolution that distinguish solar system and galactic matter. - To determine and compare the composition of the solar corona, the local interstellar medium, and GCR sources. -To search for the existence of transuranic nuclei as the signature of fresh rprocess component in CRs indicating recent nucleosynthesis occurred nearby our solar system.

94

95

-2000X2000SFM1

100 SFM2 Si-stack (AEXE)

Si-stack (AEXE)

Si-stack (AEXE)

Si-stack (AEXE)

500

SUNC

T 300 j.

SFM3

(a) ^

^

u

^

tsll HI HI H I HP ■ iH HI iH H ^

■ HBP

■ ■ ■ ■ ■ (b)

(c)

Figure 1. Particle Instrument (LTTA & SUNC) in the CORONA Experiment

B. Injection, Acceleration and Transport Processes of Energetic Particles : - To study the injection mechanism and FlP-volatility problems. - To study particle acceleration and interaction processes in solar and interplanetary ion events. - To study the evolution of CME and the relation to the particle acceleration. - To determine the time for nucleosynthesis and the time of particle acceleration in the galaxy. - To search for evidence of continuous CR acceleration by supernova shock waves. - To study the confinement or propagation mechanism of CRs in the Galaxy. - To investigate acceleration mechanism of pickup ions at the solar wind termination shock.

96

The CORONA program is a large-scaled observation program of nuclear particles with high precision. Solar, interplanetary and galactic particle events would be continuously observed on satellites or the JEM of Space Station. In order to pursue the objectives described above, we propose an instrument consisting of two kinds of telescopes, SUNC and LITA (see Fig.l). We will observe elemental composition from Z=10 to 92 in atomic number with relativistic energy, and elemental and isotopic composition H (Z=l) - Ce (Z=58) of particles in the range from a 30 MeV/n to 400 MeV/n. We will also search the existence of trans-uranium atoms in relativistic cosmic rays. We will put stress on the measurement of elemental and isotopic abundance especially for heavy elements in the GCRs and solar and interplanetary ion events, especially UH particles in space.

2

Scientific Motivation and Significance

Elemental abundance in the cosmic radiation above Z=30 has been measured by experiments on the HEAO-3, ARIEL-6 and LDEF, with earlier exploration of the Z>50 region by using plastic track detectors at high altitude balloons [3-6]. The abundance of even-Z nuclei extending from Z=30 to 60 has been individually resolved. For Zs^60, observation data are still statistically poor and the charge resolution is broadened in the HEAO and ARIEL-6 data, making it necessary to group charges for meaningful abundance measurements. Isotopic abundance heavier than Se both in the solar and galactic cosmic rays has never been measured, while the abundance of lighter elements are clearly observed especially by CRIS and SIS experiments on ACE satellite. We still lack definitive data that could provide further understanding of the origin and history of galactic matter. Here we focus and discuss only on the nuclear composition of ultraheavy CRs and the necessity of the detailed measurements for them. [A] Neutron-Rich Cosmic Ray Nuclei and s-/r-Process Nucleosynthesis The study of elemental and isotopic composition in the UH cosmic rays is important step to understand the origin and history of CRs. Nuclei of Zs£30 include contributions not only from equilibrium (e-) process and proton (p-) capture process but also from the rapid (r-) and slow (s-) neutron capture processes. The study for heavier elements, Z ^ 3 4 , will give us much opportunity of the UH observations which should delineate the r-process and the s-process contributions to the CR7). Rprocess nucleosynthesis would be expected to dominate the production of As, Se, Br, Kr, Rb, Y and Te while s-process production would dominate in Ga, Ge, Sr, Zr, Mo, Ru, Pd, Ag, Cd, In, Sn, I, Ba and Ce. Some elements are dominated by a single significant nucleosynthesis, while others are mix contributions from r-, s-, pprocesses in the nucleosynthesis.

97 According to HEAO-3 and ARIEL-6 observations, the abundance of ultraheavy nuclei was best estimated by Binns et al. (1989) [3]. Abundance in the region 32^Z60, Pb-group (81-83) is substantially deficient. Elements of 62 ^SZ^S72, Pt-group (74-80) and actinides (Z>83) are substantially high, which suggests an enhanced r-process at the GCRS. The r-process enhancement Z>60 might indicate the admixture of some unusual r-process material (not the same as in the solar system). It seems evident that these physical conditions differ from those where the solar system material was produced. There are strong reasons for the precise abundance measurements of these nuclei with high charges. A high resolution study with good statistics would give important new information on the relative contributions of r- and s-process nucleosynthesis and on the evolution of galactic matter. An accurate measurement of the U/Th ratio will give a mean age of heavy cosmic rays from the time of nucleosynthesis, whose ratio may differ from the solar value [8]. The presence of transuranic nuclei would be the signature of a fresh r-process component in the cosmic rays, indicating recent nucleosynthesis occurred nearby our solar system. Higher precision measurements of the even Z nuclei would allow more detailed comparison with solar system, special source production of cosmic rays, and r- and s-process abundance. [B] Neutron-Rich Cosmic Ray Nuclei and Wolf Rayet-Stars The Wolf Rayet model predicts that a fraction of UH nuclei originate from the material emitted from Wolf-Rayet stars [7,9,10-12]. These massive stars are undergoing significant mass loss (~10-5 solar mass per year) by high-speed stellar winds (several thousand km/s). As a result, they have been stripped of their hydrogen envelopes, and He-burning products including 12C, 1 6 0, and 22Ne and are being expelled from their surface [9,11-13]. The high-speed winds make attractive sites for the acceleration of CRs to moderate energies. Material flowing from WCstars (carbon-rich WR stars) contains gas which has been processed through core-He burning, i.e. considerably enriched into 12C, 16 0 and 22Ne. This composition is making GCR source anomalies. If massive stars contribute significantly to the 12C, 16 0, arid 22Ne excesses at the CR sources, they should also enhance other neutronrich isotopes of UH CRs. The model also predicts that 67Zn, 69Ga, 71Ga, 70Ge, 80Kr, 82Kr, and 86Sr could be enhanced including enhanced abundance of light isotopes such as 12C, 1 6 0, and 22 Ne [14]. The charge region from Z=29 to 40 is relatively free of secondary production, because there are no abundant heavier nuclei and the HEAO-3 results indicate that CR nuclei with 31S=Z^S40 observed would be primary CRs in the observed flux of nucleil [5].

98 This charge region described above presents an opportunity to identify the contributions of several nucleosynthesis processes to GCRS material. In addition, problems of acceleration and propagation mechanism of CRs will be addressed from these measurements. [C] Elemental and Isotopic Composition in Solar and Interplanetary Energetic Particles Solar system abundance, especially isotopic ones, are actually based on the terrestrial material, while meteorites serve as the standard source for elemental abundance which is characterizing solar system material. Solar energetic particles present a direct sample of solar material that is used to study the most energetic acceleration processes that occur naturally in our solar system. Comprehensive survey of solar energetic particles (SEP) have shown that the elemental composition of solar corona differs from that of the photosphere in that the abundance of elements with first ionization potential (FIP) >10eV is depleted by a factor of about 4 relative to other elements. The difference apparently indicates that neutral species are less efficiently transported from the photosphere to the corona. The measurements of the SEP isotopic composition are presently available for only elements Z2S32. The uncertainties in the existing measurements for transiron elements are still large as results of statistical limits. With a greatly improved collecting power and excellent mass resolution, it will provide the systematic exploration of the rich storage bank of solar information. And we can make much progress in our knowledge of solar isotopic composition. Multi-satellite observation of particles with other missions and ground-based observations provide key diagnostics of coronal mass ejection (CME) source regions and initiation mechanisms. Multi-satellite observation and modeling activity are focused on the connection of interplanetary features to solar events, which leads to new insights to space weather application.

99 Table 1. Characteristics of Instrument for CORONA Experiment. Two compartments in the JEM-EF would be used CORONA experiment. CORONA experiment on the JEM-EF includes two large-area telescopes, SUNC and LTTA.

1

Spectrometer for Ultraheavy Nuclear Composition (SUNC) Charge identification Trajectory system + AE-detector + Velocity-detector Scintillating Fiber Matrix Trajectory system AE-detector Si- AE-detector Velocity-detector Aerogel Cerenkov detector Charge range Ne(10) - U(92) Charge resolution < 0.5 charge unit (fwhm) >3.0 GeV/n Energy range Geometric Factor 7.9 m2sr

2

Large Isotope Telescope Array (LITA) Isotope identification The well-established AE X E algorithm Multi-module array 16 (=4X4) modules 6.6-7.3m2sr Geometric Factor One telescope module 25( = 5 X5 Si-stacks) array 1 Stack Light isotopes from H through Mg 4 Stacks Heavy isotopes from Li through Kr 20 stacks Ultraheavy Isotopes from Mg through Ce Ultraheavy Isotopes from Mg Other 15 modules through Ce Each stack Two SFMs + 7 Si-AE detectors

Characteristics of LITA/module in the CORONA Telescope-1 Telescope-2 Telescope Units 1 stacks in 14-stacks in 1Module Module Charge range Mass range Charge resolution Mass resolution Energy range GF (cm2sr)

H(l)-Mg(12) M = 1 - 26 < 0.2q (fwhm) < 0.3 amu (fwhm) 10-200MeV/n 30.6 - 42.4

Li(3) - Kr(36) M = 6-86 < 0.3q (fwhm) < 0.4 amu (fwhm) 20 -400 MeV/n 122.4-169.4

Telescope-3 20-stacks in 1Module and 15 Modules Mg(12) - Ce(58) M = 24 -142 < 0.3q (fwhm) < 0.6 amu (fwhm) 20 -600 MeV/n 66408 - 72522

100

3

Instrumental Description

In order to achieve the scientific goal, CORONA program proposes two kinds of large area telescopes, called SUNC and LITA schematically shown in Fig.la-lc. SUNC is a large area spectrometer for ultraheavy composition in relativistic GCRs and LITA is a large isotope telescope array for solar and galactic particles from H to Ce with relatively low energies from 30 MeV/n to 400 MeV/n. [A] Spectrometer for Ultraheavy Nuclear Composition (SUNC) Telescope SUNC is a large area Si-Cerenkov detector telescope (Fig.la) which measure the elemental composition of heavy cosmic ray particles with relativistic velocity, especially elements beyond the iron-peak nuclei with high statistical precision. It has a total area of 4 (= 2 x 2) m2 to measure heavy nuclei, especially ultraheavy particles with Z2?60. The geometric factor is very large, 7.9 m2sr. With a long exposure of several years (2-3 years) in space, the SUNC will measure Pt/Pb elemental region, and even U and Th. SUNC will also search for trans-uranic nuclei such as Pu, Np, and Cm. The measurement could answer the question of whether cosmic rays are freshly synthesized material recently ejected from supernovae, or simply old local galactic interstellar material accelerated indirectly by passing supernova shock wave remnants. SUNC will be a sensitive detector for the measurements of the clocks in the actinide and trans-uranic nuclear region that would be a significant contribution of recent nuclear synthesis. SUNC shown schematically in Fig. la is a scintillation fiber matrix (SFM) combined with Cerenkov detectors. SFMs, which are also used for LITA, are placed on top two and bottom for the measurement of particle trajectory. And Si-detectors used as AE-detectors measure energy losses of relativistic nuclei in the Si medium and signals from Si-detectors coincident with SFM provide triggering signals. Aerogel Cerenkov detectors are used as a threshold detector for velocity determination. The characteristics of SUNC telescope is shown in Tablel. [B] Large Isotope Telescope Array (LITA) for Ultraheavy Cosmic Ray Observation The schematic configuration of the telescope (Large Isotope Telescope Array, LITA) is shown in Fig.la, Fig.lb and lc. LITA is a multi-module array (16 (= 4X4) modules : a module = 5X5 Si-stack array). It has a geometric factor of 6.65 - 7.27 m2sr in total which depends on the energy and incident species. In a telescope module, there are 25 Si-detector stack (see Fig.lb). One stack in one module is to measure light isotopes starting from H (Z=l) through Mg (Z=14). 4 stacks in the module measure isotopes from Li (Z=3) to Kr (Z=36), and other 20 stacks are for C (Z=6) - Ce (Z=56) isotopes. The remaining 15 modules measure heavy isotopes starting C (Z=6) through Ce (Z=58). Each stack consists of SFM and 7 AE detectors

101

in each unit (see Fig. lc). Measurements of particle trajectories are made by the use of SFM providing two-dimensional coordinates with an resolution (fwhm) of 0.5. SFM used for LITA are common to those for SUNC. Isotope identification can be made by the well-established AE X E algorithm. The characteristics of LITA telescope is shown in Tablel. [C] Electronics for CORONA Program The electronics box placed under bottom SFM in SUNC consists of digital processing circuits, CPU, low and high voltage supply and housekeeping monitor circuits. Analog electronics are installed nearby their detectors. The energy loss of particles entering the telescope is precisely measured by pulse height analyzers. The data produced by the SUNC/LITA contain pulse height and trajectory measurement for individual particles. CPU in each module categorizes the particle events to assign priority and optimize the mix of events. A small fraction of the data stream is used for housekeeping measurements. A valid particle event must penetrate at least top SFM and reach or penetrate the second SFM in the stack. For such events, the CPU reads a 14-bit pulse height from all of the Si-detectors including SFMs. To maximize the number of events sent to the ground, event data are compressed onboard. In addition to the data compression, diagnostic and calibration data will be sent back. The box is surrounded with by aluminum plate providing an RF shield around the module. CORONA instrument is covered with thermal blankets. CORONA's field of view is so wide that full FOV would be desired to be completely unobstructed. The structure should be designed to endure vibration and acoustic environment during launch.

4

Expected Performance

Species observed by CORONA experiment are ranging wide from H to U. The energy interval to be measured by the SUNC and LITA depends on elements. Some stacks used in the LITA experiment also measures hydrogen, helium and CNO flux which provide us the baseline in various particle events. Those fluxes are used as a baseline to examine the correlation with the fluxes of ultraheavy nuclei. The energy to be measured by SUNC and LITA is shown in Table 2 together with nuclear charge and mass resolutions expected. Clear separation of isotope masses for ultraheavy nuclei (Z2a50) in the GCRs is expected to be less than 0.20.25 amu (rms) by this Si-detector telescope array. And nuclear charge even for Pt and Pb groups will be clearly identified with about 0.2 charge unit (rms). SUNC has extremely large geometric factors. Total value of GF for SUNC becomes 7.9 m2 sr. It collects 2.0X108 Fe, 1100 Pt-group, 330 Pb with relativistic

102

energies for 3 years. A few year observation enables us to detect actinide elements and possibly to detect trans-uranium elements. Table 2 shows the number of ion events expected from SUNC for elements Fe to U [19]. LITA resolves isotopes starting from H isotopes up to at least M~110-120, and also abundant isotopes of heavier nuclei. The collecting power of such telescopes depends on the energy and species. The geometric factors GF for LITA are 30.6 - 42.4 cm2 sr for H -Mg isotopes, 122 -170 cm2 sr for Li - Kr isotopes and 6.6 -7.3 m2 sr for heavy isotopes which corresponds the geometric factor increased more than 300 times that for the GEOTAIL [16-18]. Approximate numbers for light nuclei of C with an integration four stacks for one year period of solar minimum to be collected by LITA will be >1.5 x 106. Approximate numbers for UH nuclei greater than Fe (Z=26) with an integration seven modules for 3 year period of solar minimum will be >2 x 106.

5

Necessity of Large-scaled Spacecraft Platform in Cosmic Ray Observation

The best location for performing this kind of measurements is outside the Earth magnetosphere to give the maximum number of particles. A polar orbiter of the nearEarth would be possible but the efficiency for collecting particles would become small. Because of the orbit of the Space Station which resides inside the Earth magnetosphere, rigidity cutoff effect decreases the intensity of low energy particles by about a few % compared to the outside of the Earth magnetosphere. Here we describe the necessity of cosmic ray observation using the polar orbiter or orbiter with high inclination. The intensity of heavy elements, especially trans-iron species in the cosmic ray, is extremely small so that the enlargement of the telescope and continuous observations for a long term are indispensable. Heavy element observations are not achieved on the ground. Only by using the large-scale satellites or the space station, meaningful results can be established. The advantages in the observation using the polar orbiter are: - Continuous observation over a few years - Small contamination of secondary cosmic rays from the atmosphere - Large and relatively heavy scientific payload acceptable - Large electric power available - High telemetry Particle telescope proposed in the CORONA program will accomplish a largescaled observation with high accuracy and high sensitivity that has never been realized before. The large-scale isotope observation has not been proposed elsewhere in the world. Relativistic particle observation up to the knee-energy

103

region is adopted as the ACCESS/KLEM-3 project using Space Station base, but the isotope observation is not included. The ENTICE program [20] is quite similar to the CORONA one except for the isotope measurement. ECCO would be dedicated to the observation of UH (Z>70) nuclei in GCR using the large-area glass detectors. ENTICE and ECCO would join together and a new experiment called HNX will start using space shuttle. Those projects have been proposed as a second-period observation plan of the Space Station experimental module. Therefore, it is important to execute the CORONA experiment in the same period or earlier than that the ACCESS/HNX and KLEM experiments. From this point of view, it is essential to achieve the CORONA program in the second-period observation time at least of the JEM-EF, or another choice is to collaborate a joint experiment with Russia and/or US using large satellite platform. International cooperative system with Russia/US and other countries will be a great help to promote researches and make a new progress in cosmic ray research. 6

Concluding Remark

A high quality and continuing series of observations of solar and galactic cosmic rays provide important exploratory data. Definitive measurements are now close to the availability of spacecraft capable of carrying very large scientific payloads for long extended period of years. Scientific objectives in such measurements are to determine the composition of rare ultraheavy nuclei up to uranium and trans-uranium at high energies and isotopic abundance above iron-peak elements. In these research fields, we have no date at all or still lack of decisive data that could characterize the origin of cosmic rays, the nucleosynthesis processes leading to cosmic ray matter and governing the chemical evolution of our galactic matter. And we are also short of definitive data that could further provide the understanding of the structure and properties of the interstellar medium, and the acceleration and transport mechanism of cosmic ray particles in the galaxy. In order to achieve the objectives, a large spacecraft platform and/or the International Space Station are best suited in the multilateral cooperation with Japan and other countries. The CORONA program is the large-scaled observation program of nuclear particles with high precision. Continuous observations of solar, interplanetary and galactic particle events would be observed on exposure facility in space like the JEM of International Space Station, Shuttle experiment and or Russian satellites with international collaboration with Russia and other countries. The development team for the CORONA Program is at present organized from core members in GEOTAIL HEP team and some other scientists from AoyamaGakuin Univ., Chiba Univ., and National Institute for Radiological Science, RIKEN, and SNI of Moscow State University as key institute as the Russian counterpart.

104

7

Acknowledgements

We appreciate the cooperative work with Dr. Shibata at Aoyama-Gakuin University, Dr. Uchihori at NIRS, Dr. Takashima at Nagoya University, Dr. Tanihata at RIKEN, and Dr. Kawai at Chiba University in the development of detectors. References 1. N. Hasebe, Cosmic Radiation, 1(1999)15-30 in Japanese. 2. N. Hasebe, Doke-Symposium. 3. W.R. Binns. et al., Proc. AIP Conf., Cosmic Abundance of Matter, 183(1989a)147. 4. W.R. Binns. et al., Astrophys. J., 297(1985)111. 5. W.R. Binns. et al., Astrophys. J., 346(1989)997. 6. P.H. Fowler, Astrophys. J., 314(1987)739. 7. J.P. Meyer, Proc. 17th Int. Cosmic Ray Conf. (Paris), 2(1981)3265. 8. A.J. Westphal et al., Proc. 24th Int. Cosmic Ray Conf. (Rome) 2(1995)581. 9. J.P. Meyer, L.O'C. Drury and D. Ellison, Astrophys. J., 487(1997)182. 10. M. Casse and J.A. Paul, Astrophys. J., 258(1982) 860. 11. A. Maeder, Astron. Astrophys., 120(1983)130. 12. N. Prantoz, M. Arnold and J.P. ArCoragi, Astrophy. J., 315(1987)209. 13. J.P. Meyer, Proc. 19th Int. Cosmic Ray Conf. (Calgary), 2(1985)141. 14. R.A. Mewaldt, Proc. AIP Conf., Particle Astrophysics, 203 (1990)268. 15. N. Prantoz et al., Proc.l9th Int. Cosmic Ray Conf. (Calgary), 3(1985)167. 16. T. Doke et al., J. Geomag. and Geophys., 46(1994)713. 17. N. Hasebe et al., Jpn. J. Appl. Phys., 31(1992)1191. 18. N. Hasebe et al., Nucl. Instrum. and Methods, Phys. Res.,A325(1993)335. 19. W.R. Binns. et al., Proc. AIP Conf., Particle Astrophysics, 203(1990)231. 20. W.R. Binns. et al., Proc. 25th Int. Cosmic Ray Conf. (Durham) 1(1997)65.

III. Stellar Evolution and the Nucleosynthesis - Hydrostatic Burning -

This page is intentionally left blank

Direct Capture S-factors from Asymptotic Normalization Coefficients R.E. Tribble, A. Azhari, H.L. Clark, C.A. Gagliardi, Y.-W. Lui, A.M. Mukhamedzhanov, A. Sattarov, X. Tang, L. Trache Cyclotron Institute, Texas A&M University, College Station, Texas 77843 V. Burjan, J. Cejpek, V. Kroha, S. Piskof, J. Vincour Institute for Nuclear Physics, Czech Academy of Sciences, Prague-Rez, Republic F. Carstoiu Institute for Atomic Physics, Bucharest,

Czech

Romania

Peripheral transfer reactions can be used to determine asymptotic normalization coefficients (ANC). These coefficients, which provide the normalization of the tail of the overlap function, determine S-factors for direct capture reactions at astrophysical energies. A variety of proton transfer reactions have been used to measure ANC's. As a test of the technique, the 16 0( 3 He,d) 17 F reaction has been used to determine ANC's for transitions to the ground and first excited states of 1 7 F. The S-factors for le O(p,7) 17 F calculated from these 17 F -+ 1 6 0 + p ANC's are found to be in very good agreement with recent measurements. Following the same tech­ nique, the 10 B( 7 Be, 8 B) 9 Be and 14 N( 7 Be, 8 B) 13 C reactions have been used, along with optical model parameters for the radioactive beams that were obtained from a study of elastic scattering of loosely bound p-shell nuclei, to measure the ANC ap­ propriate for determining 7 Be(p,7) 8 B. The results from the two transfer reactions provide an indirect determination of S 17(0).

1

Introduction

Stellar evolution involves sequences of capture reactions and beta decays. Pre­ dicting the evolution of a star requires knowing reaction rates and half lives. Direct capture reactions of astrophysical interest usually involve systems where the binding energy of the captured proton is low. Hence at stellar energies, the capture proceeds through the tail of the nuclear overlap function. The shape of the overlap function in this tail region is completely determined by the Coulomb interaction, so the amplitude of the overlap function alone dic­ tates the rate of the capture reaction. The 7 Be(p,7) 8 B reaction is an excellent example of such a direct capture process. Indeed recent calculations of the nor­ malization constant have been used to predict the capture rate 1'2. But new measurements, both direct and indirect, are still needed as was underscored in a recent review of stellar reaction rates 3 which includes a detailed discussion of the uncertainties in our present knowledge of Si7(0) and its importance to

107

108

the solar neutrino problem. The asymptotic normalization coefficient (ANC) C for A+p ++ B specifies the amplitude of the tail of the overlap function for the system. In previous communications 1 ' 4 , we have pointed out that astrophysical S-factors for pe­ ripheral direct radiative capture reactions can be determined through measure­ ments of ANC's using traditional nuclear reactions such as peripheral nucleon transfer. Direct capture S-factors derived with this technique are most reliable at the lowest incident energies in the capture reaction, precisely where capture cross sections are smallest and most difficult to measure directly. Of course it is extremely important to test the reliability of the technique in order to know the precision with which it can be applied. Determining the S-factors for 1 6 0(p,7) 1 7 F from its ANC's has been recognized as a suitable test for this method 3 because the results can be compared to existing direct measurements of the cross sections 5,6 . Furthermore, the 1 6 0(p,7) 1 7 F reaction has substantial similarities to the 7 Be(p,7) 8 B reaction. As part of an ongoing program to mea­ sure ANC's, we have used the proton exchange reactions 9 Be( 10 B, 9 Be) 10 B and 13 C( 1 4 N, 1 3 C) 1 4 N to measure the ANC's for 10 B -> 9 Be + p and 14 N ->• 13 C + p. Below we briefly summarize these results. This is followed by a discus­ sion of a measurement of the another proton transfer reaction, 1 6 0( 3 He,d) 1 7 F, which is used to determine the ANC's for the ground and first excited states in 1 7 F . From these ANC's, we calculate S-factors for both 9 Be(p,7) 10 B and 16 0(p,7) 1 7 F and compare to experimental results. Finally we discuss our mea­ surement of the 10 B( 7 Be, 8 B) 9 Be and 14 N( 7 Be, 8 B) 13 C reactions, the extraction of the ANC's for 8 B -»• 7 Be + p and our determination of Si7(0).

2

A N C ' s from Proton Transfer Reactions

Traditionally spectroscopic factors have been obtained from proton transfer reactions by comparing experimental cross sections to DWBA predictions. For peripheral transfer, we show below that the ANC is better determined and is the more natural quantity to extract. Consider the proton transfer reaction a + A -» c + B, where a = c + p, B = A + p. As was previously shown 7 we can write the DWBA cross section in the form &1 _ V^ (CAplBjB) jBja

AplBJB

iCcpUjJ

~DW

m

Cplaja

where aP^f , „• is the reduced DWBA cross section and ji,U are the total and orbital angular momenta of the transferred proton in nucleus i. The factors bcpUja and bApiBjB a r e t n e ANC's of the bound state proton wave functions in

109

nuclei a and B. If the reaction under consideration is peripheral, the ratio frDW T>

_

iBJBlgja

AplBjB

(n\

°cplaja

is independent of the single particle ANC's bApiBjB a n ( i bcptaja. Thus for surface reactions the DWBA cross section is best parametrized in terms of the product of the square of the ANC's of the initial and final nuclei ( C B ) 2 ( C a ) 2 rather than spectroscopic factors. We have used this formulation to extract ANC's from the peripheral pro­ ton transfer reactions 9 Be( 10 B, 9 Be) 10 B, 13 C( 14 N, 13 C) 14 N and 1 6 0( 3 He,d) 1 7 F. The first two reaction studies were carried out with beams of 10 B and 14 N from the K500 superconducting cyclotron at Texas A&M University. Both elastic scattering and transfer reaction products were measured in the MDM spectrometer. Details of the experiments can be found in 7 ' 8 , including the re­ sults of DWBA fits using the code PTOLEMY 9 with optical model parameters obtained from an analysis of the elastic scattering and the extracted ANC's. The 1 6 0( 3 He,d) 1 7 F reaction was measured previously at a beam energy of 25 MeV 10 . We repeated the measurement at 29.75 MeV in order to obtain better angular coverage and to have a measurement at a second energy, both of which were necessary for extracting reliable ANC's. Data at laboratory scat­ tering angles between 6.5° and 25° were obtained using Si solid state detectors and a 3 He beam, incident on a 134 /zg/cm2 Mylar target, from the U-120M isochronous cyclotron of the Nuclear Physics Institute of the Czech Academy of Sciences. Additional data at laboratory angles between 1° and 11° were obtained using the MDM magnetic spectrometer and a molecular (3He—d)+ beam, incident on a 540 fig/cm2 Mylar target, from the Texas A&M University K500 superconducting cyclotron. Absolute cross sections were determined at the NPI using their detection system which has been well calibrated for (3He,d) reaction studies. The data obtained at TAMU were normalized to the data from the NPI measurement in the region where the two data sets overlapped. More details of the experiments can be found in 1 1 . In order to extract ANC's, DWBA calculations were carried out with the finite range code PTOLEMY, using the full transition operator. A check on the extracted ANC's versus Woods-Saxon well radial parameters indicated that the calculated DWBA cross sections are insensitive to assumptions about the 17 F wave functions in the nuclear interior. A range of optical model parameter sets was studied for both the entrance and exit channels, as detailed i n 1 1 . Normalizing the DWBA calculations to the data and dividing by the ANC's for the single particle orbitals yields the product of the ANC's for the 1 7 F -»■ 1 6 0 + p and 3 He -> d + p systems. Dividing this product by the known ANC for 3 He

110

-»• d 4- p 1 2 ' 1 3 provides C 2 for 1 7 F -»• 1 6 0 + p. The dominant contribution to the uncertainties is due to the variation in the extracted ANC's with different optical model parameter sets. Our final adopted ANC's are C 2 = 1.08(10) fm _ 1 and 6490(680) fin-1 for the ground and excited states, respectively. 3

Using A N C ' s to Predict Astrophysical S-Eactors: Test Cases

The ANC's found from the proton transfer reactions can be used to determine direct capture rates at astrophysical energies. Astrophysical S-factors have been determined for both 9 Be(p,7) 10 B and 1 6 0(p,7) 1 7 F as tests of the tech­ nique. The relation of the ANC's to the direct capture rate at low energies is straightforward to obtain. The cross section for the direct capture reaction A + p -> B + 7 can be written as a = A||2,

(3)

where A contains kinematical factors, I%p is the overlap function for B -> A+p, O is the electromagnetic transition operator, and ip\ ' is the scattering wave in the incident channel. If the dominant contribution to the matrix element comes from outside the nuclear radius, the overlap function may be replaced by W 2Kr)

lUr)*C ->»f

,

(4)

where C defines the amplitude of the tail of the radial overlap function I%p, W is the Whittaker function, r\ is the Coulomb parameter for the bound state B = A+p, and n is the bound state wave number. The required C's are just the ANC's found above from transfer reactions. Thus, the direct capture cross sections are directly proportional to the squares of these ANC's. Using the results outlined above, the S-factors describing the capture to both the ground and first excited states for 1 6 0(p,7) 1 7 F were calculated, with no additional normalization constants. The results are shown in Fig. 1 com­ pared to the two previous measurements of 1 6 0 ( p , 7 ) 1 7 F 5 ' 6 . Both E l and E2 contributions have been included in the calculations, but the E l components dominate the results. The theoretical uncertainty in the S-factors is less than 2% for energies below 1 MeV. Above 1 MeV the nuclear interaction begins to be important in the evaluation of the scattering wave function. The agreement between the measured S-factors and those calculated from our 1 7 F —> 1 6 0 + p ANC's is quite good, especially for energies below 1 MeV where the ap­ proximation of ignoring contributions from the nuclear interior should be very

111

Ep(MeV) Figure 1: A comparison of the experimental S-factors to those determined from the ANC's found in 1 6 0 ( 3 H e , d ) 1 7 F . The solid data points are from 5 , and the open boxes are f r o m 6 . The solid lines indicate our calculated S-factors, and the dashed lines indicate the ±1 7 Be + p. The radioactive 7 Be beam was produced at 12 MeV/u by filtering reaction products from the 1 H( 7 Li, 7 Be)n reaction in the recoil spectrometer MARS, starting with a primary 7 Li beam at 18.6 MeV/A from the TAMU K500 cy­ clotron. The beam was incident on an H 2 cryogenic gas target, cooled by LN 2 , which was kept at 1 atm. (absolute) pressure. Reaction products were mea­ sured by 5 cm x 5 cm Si detector telescopes consisting of a 100 fan AE strip detector, with 16 position sensitive strips, followed by a 1000 pim E counter. A single 1000 /xm Si strip detector was used for initial beam tuning. This detector, which was inserted at the target location, allowed us to optimize the beam shape and to normalize the 7 Be flux relative to a Faraday cup that measured the intensity of the primary 7 Li beam. Following optimization, the approximate 7 Be beam size was 6 mm x 3 mm (FWHM), the energy spread was « 1.5 MeV, the full angular spread was A# « 28 mrad and A 0 « 62 mrad, and the purity was >99.5% 7 Be for the experiment with the 1 0 B target. The beam size and angular spread were improved for the experiment with the

113 14

N target to 4 mm x 3 mm (FWHM), A0 « 28 mrad and A<j> » 49 mrad. Periodically during the data acquisition, the beam detector was inserted to check the stability of the secondary beam tune. The system was found to be quite stable over the course of the experiment with maximum changes in intensity observed to be less than 5%. The typical rate for 7 Be was « 1.5 kHz/pnA of primary beam on the production target. Primary beam intensities of up to 80 pnA were obtained on the gas cell target during the experiments. In order to extract ANC's from reactions involving radioactive-ion beams, it is necessary to have optical model parameters. Typically radioactive beam intensities are too small to measure elastic scattering and obtain good optical model parameters. Consequently we have carried out a series of elastic scat­ tering measurements with stable beam and target combinations that are close to those for our radioactive beam measurements. We used the folding model prescription to calculate the potential parameters and then renormalized the real and imaginary parts of the potentials to fit the data. Several different in­ teractions were tried but the most consistent results were found to be from the JLM interaction 19 . The renormalization coefficients were found to be 0.366 ± 0.014 for the real potential and 1.000 ± 0.087 for the imaginary potential. The uncertainties listed were from the dispersion in the renormalization coefficients that we found for the different reactions 20 . The large renormalization for the real part of the optical potential is mostly due to dynamic polarization effects which are not fully accounted for in the folding model. With these poten­ tial parameters, it is now possible to calculate elastic scattering and transfer reactions for loosely bound p-shell systems. Elastic scattering angular distributions for 7 Be on the 1 0 B and 14 N targets are shown in Fig. 3. For the 1 0 B target, the elastic scattering yield includes con­ tributions from three target components, 10 B(86%), 12 C(10%) and 1 6 0(4%), while the Melamine target includes 14 N(67%), 12C(28%) and ^ ( 5 % ) . A Monte Carlo simulation described below was used to generate the solid angle factor for each angular bin and the smoothing needed for the calculation to account for the finite angular resolution of the beam. The absolute cross section is then fixed by the target thickness, number of incident 7 Be, the yield in each bin, and the solid angle. The curves shown with the elastic scattering were found from the optical model, with the parameters discussed above, by adding together the cross section predictions for the target components in the labo­ ratory frame and then transforming the result to the center of mass assuming kinematics appropriate for either the 1 0 B or 14 N targets. In both cases, the optical model calculations are compared to the data without additional nor­ malization coefficients. The detector resolution is not sufficient to distinguish inelastic excitations from elastic scattering. This likely explains why the data

114

10

15

20

25

30

35

tfc.m. ( d e g )

Figure 3: Angular distributions for elastic scattering from the 10 B and 14N targets. The dashed curves are from optical model calculations of the target components and the solid curves are smoothed over the angular acceptance of each bin.

exceed the calculations in the minima. Overall, the agreement between the measured absolute cross sections and the optical model predictions is excellent thus providing confidence that our normalization procedure is correct. a

10

The 8 B Q-value spectra, shown in Fig. 4, were obtained by assuming either B( 7 Be, 8 B) 9 Be or 14 N( 7 Be, 8 B) 13 C reaction and correcting the 8 B reaction

115

-14

-12

-8

-10

-6

Q (MeV) 200

-14

-12

-10

Q (MeV) Figure 4: Q-value spectra for 8 B reaction products on the 10 B (top panel) and Melamine targets (bottom panel). The three peaks in the 10 B target spectrum correspond to the excitation of the ground state and second excited states of 9 Be and the ground state in 18 N from the l e O contamination in the target. Data for the Melamine target show a clear isolation of the ground state for 1 3 C.

products for kinematic energy shifts as a function of scattering angle. In the case of the 10 B target, the major contributions to the energy resolution are the beam energy spread, the target thickness and the nonuniformity of the target. The beam energy spread and differential energy loss in the target dominated the energy resolution for the Melamine target. Since the ground state of 9 Be is not cleanly separated from excited states, a Monte Carlo simulation of the experiment has been used to fix the line shape and determine cross sections.

116

The simulation, which is fine tuned to reproduce the measured beam properties and the resolution observed in elastic scattering, includes the geometry of the experimental setup, reaction kinematics, nonuniform energy loss in the target and the size, angular spread and energy spread of the beam. The beam location and angle at the target are determined by symmetry requirements on 7 Be elastic scattering data. The three peaks shown in Fig. 4 for the 1 0 B target correspond to the excitation of the ground and second excited states of 9 Be and the ground state of 15 N from the 1 6 0 contamination in the target. They were obtained by including the predicted angular distributions for the states in the Monte Carlo simulation and then extracting the associated Q-value spectrum. The normalization of the three peaks was done by a x2 minimization to the data. The cross section ratio for transitions to the ground and second excited states in 9 Be is in good agreement with theoretical expectations 21 . In the Melamine case, the ground state of 13 C is cleanly resolved from excited states making the normalization of the Q-value spectrum via the Monte Carlo simulation straight forward. The ANC for 8 B -> 7 Be + p was extracted based on the fit to the Q-value spectra and the ANC's 7 ' 8 for 1 0 B ->■ 9 Be + p and 14 N -> 13 C + p following the procedure outlined above in our test case. Two 8 B orbitals, lpi/2 and 1P3/2) contribute to the transfer reaction but the lp 3 /2 dominates in both cases. In calculating the angular distributions, we used the ratio for the two orbitals as given by a microscopic description of the 8 B ground s t a t e 4 . The optical model parameters were obtained from renormalized microscopic folding potentials using the JLM effective NN interaction 19 described above. The entrance channel parameters were the same as those used in calculating the elastic scattering angular distributions for 7 Be on 1 0 B and 14 N in Fig. 3. We have checked the sensitivity of the calculations by varying the normalization parameters. As in previous studies, the results are insensitive to bound state single particle well parameters in the DWBA calculations. Angular distributions for the ( 7 Be, 8 B) reactions populating the ground states of 9 Be and 13 C were extracted using the same procedure as for the elastic scattering. The results are compared to DWBA calculations in Fig. 5. The normalization factors between the data and calculations were obtained from the fits to the respective Q-value spectra. The astrophysical S-factor for 7 Be(p,7) 8 B has been determined from the ANC which includes an 8.1% uncertainty for optical model parameters, an uncertainty for experimental fits and normalization of the absolute cross section of 10.9% for the 10 B target and 8.1% for the 14 N target and the uncertainty in the ANC's for 1 0 B -> 9 Be + p and 14 N -» 13 C + p. The relative contribution of the two angular momentum couplings to the S-factor is straightforward to

117

10

15

20

25

30

tf=.m. ( d e g )

5

10

t5

20

^.m. ( d e g ) Figure 5: Angular distributions for 8 B populating the ground state of 9 Be from the 10 B target and I 3 C from the Melamine target. In the top figure, the dashed curve shows the result of a DWBA calculation for the dominant component that contributes to the cross section. Two components are shown in the bottom figure. In both cases, the solid curve is smoothed over the angular acceptance of each bin.

calculate and introduces a negligible additional uncertainty in our result 1>4 . The values that we find are Si7(0) = 18.4 ± 2.5 eV b for the 1 0 B target and 16.6 ± 1.9 eV b for the 14 N target, which are in good agreement with the recommended value 3 of 19J]4, eV b. One of the primary sources of uncertainty in the values quoted above for Si7(0) is the optical model parameters that are used to predict the angular

118

distribution. As indicated, we have developed a set of global optical model parameters for use with radioactive beams in this mass and energy region. Since the optical model parameters for the two different targets are derived by the same technique, this introduces correlations in the uncertainties between the two results. Accounting for this correlation, we find the S-factor from the combined measurements to be 17.2 ± 1.8 eV b. Recently we completed a measurement of the 1 4 N( 1 1 C, 1 2 N) 1 3 C reaction with a U C beam at about 110 MeV. Following the analysis of these data we will determine the direct capture rate for the 11 C(p,7) 12 N reaction which is important for stellar burning in low metallicity massive stars. This work was supported in part by the U.S. Department of Energy under Grant number DE-FG05-93ER40773 and by the Robert A. Welch Foundation. References 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21.

H.M. Xu et al, Phys. Rev. Lett.73, 2027 (1994). L.V.Grigorenko et al, Phys. Rev. C 57, R2099 (1998). E.G. Adelberger et al, Rev. Mod. Phys. Vol. 70(4), 1265 (1998). A.M. Mukhamedzhanov and N.K. Timofeyuk, J E T P Lett. 51, 282 (1990). R. Morlock et al, Phys. Rev. Lett. 79, 3837 (1997). H.C. Chow, G.M. Griffith and T.H. Hall, Can. J. Phys. 53,1672 (1975). A.M. Mukhamedzhanov et al., Phys. Rev. C 56, 1302 (1997). L. Trache et al, Phys. Rev. C 58, 2715 (1998). M. Rhoades-Brown, M. McFarlaneand S. Pieper, Phys. Rev. C 2 1 , 2417 (1980); Phys. Rev. C 2 1 , 2436 (1980). J. Vernotte et al, Nucl. Phys. A571, 1 (1994). C.A. Gagliardi et al, Phys. Rev. C 59, 1149 (1999). M. Kamimura and H. Kameyama, Nucl. Phys. A508, 17c (1990). A.M. Mukhamedzhanov, R.E. Tribble and N.K. Timofeyuk, Phys. Rev. C 51, 3472 (1995). R. Morlock, private communication. A. Sattarov et a/., Phys. Rev. C 60, 035801 (1999). D. Zhanow et al, Nucl. Phys. A589, 95 (1995). A. Azhari et al, Phys. Rev. Lett. 82, 3960 (1999). A. Azhari et al, Phys. Rev. C 60, 035801 (1999). J.P. Jeukenne, A. Lejeune and C. Mahaux, Phys.Rev. C 16, 80 (1977). L. Trache et al, Phys. Rev. C, in press. S. Cohen and D. Kurath, Nucl. Phys. 73, 1 (1965).

SOLAR NEUTRINO PROBLEM RELATED NUCLEAR PHYSICS EXPERIMENTS

WEIPING LIU'' 5 , XIXIANG BAI', SHUHUA ZHOU', ZHANWEN MA U - ZHICHANG LI 1 , YOUBAO W A N G ' 3 , ANLI LI', ZHONGYU MA', BAOQIU CHEN', XIAODONG TANG'- 4 , YINLU HAN', QINGBIAO SHEN' AND JINCHENG XU' 'China Institute of Atomic Energy, P.O. Box 275(46), Beijing 102413, P. R. China E-mail: wpliu@iris. ciae. ac. en 2

3

Department

Department

of Physics and Astronomy, 401 Nielsen Physics Building, The University of Tennessee, Knoxville, Tennessee 37996-1200, U.S.A

of Physics, University ofJyvaskyla,

P.O. Box 35 (Y5), 40351, Jyvaskyla,

Finland

4

Cyclotron Institute, Texas A&M University, College Station, Texas 77843, U.S.A

M. HELLSTROM 5 - 6 , R. COLLATZ 6 , J. BENLLIURE 6 , L. CHULKOV 7, D. CORTINA GIL 6 , F. FARGET 6 , H. GRAWE 6 , Z. HU 6 , N. IWASA 6 - 8 , M. PFUTZNER6-9, A. PIECHACZEK 1 0 , R. RAABE 1 0 ,1. REUSEN 10, E. ROECKL 6 , G. VANCRAEYNEST' 0 , A. WOHR 10 5

Lund University, Department of Physics, Division of Cosmic and Subatomic Physics, P.O. Box 118, S-22100 Lund, SWEDEN 6

Gesellschaftfur

Schwerionenforschung,

7

D-64291 Darmstadt,

Kurchatov Institute, 123182 Moscow, Russian

8

Germany

Federation

RIKEN, Institute of Physical and Chemical Research, Saitama 351-01,

9

Institute of Experimental

Japan

Physics, University of Warsaw, PL-00681 Warsaw,

'"Institut voor Kern-en Stralingsfysika,

Katholieke Universiteit, B-3030 Leuven,

Poland Belgium

The two measurements, performed in connection with the solar neutrino problem, are described. The 7Be(d,n)8B reaction was measured at E cm = 5.8 and 8.3 MeV for deducing the Sn(0) factor for the 7Be(p,y)8B reaction. The angular distribution data were analyzed by using a distorted-wave Born approximation. The S,7(0) factor for the 7Be(p,Y)8B reaction was derived to be 27.4±4.4 eV b through the asymptotic normalization constant extracted from the experimental data. The (3-decay of 4200 0)

(a)

= 150

4

0

iifioo ;o o ■§50 X

*1*"«

/

(b)

> 0.4

i

i

Erel [MeV] Figure 2: (a) The p- 7 Be coincidence yields plotted as a function of relative energy. The solid and dashed histograms denote simulated E l + M l and E l yields, respectively, (b) the p- Be coincidence efficiency calculated by Monte Carlo simulations.

layers of matter was taken into account in the simulation. Further corrections in the simulation are due to the feeding of the excited state at 429 keV in 7 Be. We used the result by Kikuchi et al}1 who measured the 7-decay in coincidence with the CD of 8 B . The histograms in Fig. 2(a) show the simulated E l + M l yields (solid) and El yields (dashed). As seen in this figure, the shape and magnitude of the experimental energy dependence are fairly well reproduced. This indicates that the CD yield is well described by the combination of the Ml resonance and the pure El continuum. The total efficiency calculated by the simulations was shown in Fig. 2 (b) which is high over the entire Ere\ range covered in our study. The relative-energy resolution was estimated from the simulation to be e.g. 2p+a: reaction at the experimental area of RCNP Osaka, in the energy region from lOOkeV. ACKNOWLEDGEMENT This work was supported by the Grant-in-Aid of Scientific Research, Ministry of Education, Science, Culture and Sports.

1. J. N. Bahcall, W. F. Huebner, S. H. Lubow, P. D. Parker, and R. K. Ulrich Rev. Mod. Phys. 54, 767 (1982). 2. M. Junker, A. D'Alessandro, S. Zavatarelli, C. Arpesella et al , Phys. Rev. C57, 2700 (1998) and C. Arpesella et al Phys. Lett. B389, 452

145

2S0

500

750

1000

12S0

1500

1750

2000 2250 250O cnergy(channcl)

Figure 4: Energy spectra of d + 3 H e reaction for E-count(upper) and AE-counter(lower).

and U. Greife et al., Nucl. Instr. and Meth. A350, 327 (1994). 3. H. Ohsumi, H. Ejiri, M. Fujiwara, K. Fushimi, K. Hayashi, R. Hazamza, T. Kishimoyo, M. Komori, N. Kudomi, K. Kume, K. Matsuoka, H. Miyazaki, T. Nitta, N. Suzuki, and S. Tasaka, RCNP Annual Report, RCNP 1995, p. 175. 4. T. Itahashi, K. Takahisa, N Kudomi M. Komori, et.al., Rev. Sci. Instr. 69, 1032 (1998) ; T. Itahashi, T. Iki, M. Komori et. al., The 11th Symp. on Ace. Sci. and Tech., Harima, Hyogo, Japan, p. 45.; T. Itahashi, N Kudomi K. Kume, M. Komori, et.al., Rev. Sci. Instr. (2000) to be published. 5. P. Sortais, C. Bieth, P. Foury, et. al., Proc. of the 12th Intern. Workshop on ECR ion sources, RIKEN, Japan 1995 p. 45. 6. K. Takahisa et al, 24th INS Symposium on ECR Ion sources and their Applications, April 25-27, RIKEN 1995 p.115, and RCNP Annual Report, RCNP 1994 p.190. 7. G. Fiorentini, R. W. Kavanagh, C. Rolfs, Z. Phys. A350, 289 (1995). 8. K. Laganke, T.D. Shoppa, C. A. Barnes, C. Rolfs, Phys. Letts. B369, 211 (1996). 9. U. Greife, F. Gorris, M. Junker, C. Rolfs, and D. Zahnow, Z. Phys. A351, 107 (1995). 10. Proc. of the 19th INS Symp. Cooler Rings and Their Applications, Tokyo, Japan 1990

This page is intentionally left blank

IV. Nucleosynthesis in Explosive Burning and New Approach

This page is intentionally left blank

PROBING STELLAR EXPLOSIONS WITH RADIOACTIVE BEAMS AT ORNL

MICHAEL S. SMITH Physics Division, Oak Ridge National Laboratory, MS-6354, Bldg. 6010, P.O. Box 2008, Oak Ridge, TN 37831-6371, USA E-mail: [email protected]. ornl.gov Measurements of nuclear reactions on radioactive isotopes are necessary to understand stellar explosions such as novae and X-ray bursts. We have recently measured the l7F(p,p)"F, l7 F(p,a) l4 0 ; l8F(p,p)l8F, and '"F(p,a)150 excitation functions with beams of radioactive ions produced at the Holifield Radioactive Ion Beam Facility (HRIBF) at Oak Ridge National Laboratory. Our experimental setup includes a Silicon Detector Array and the Daresbury Recoil Separator coupled to a windowless, differentially-pumped hydrogen gas target system. We have also measured the l2C(p,y)L1N cross section to commission our recoil separator for proton capture reactions on radioactive isotopes. To support these measurements, we are making unique calculations of isotope synthesis in stellar explosions to determine the effect of nuclear physics uncertainties on explosion model predictions. We are also making detailed evaluations of some nuclear reactions rates that are important input for explosion models.

1

Physics Motivation

There are a number of important astrophysical events during which hydrogen serves as fuel for (i.e., is burned by) (p,y) fusion reactions under non-hydrostatic equilibrium conditions. These explosive hydrogen burning events, which include novae and X-ray bursts, are among the most energetic explosions (~1038 - 1045 ergs) known in the universe. Furthermore, these explosions affect the evolution of binary star systems and synthesize some of the elements comprising our bodies and the world around us. For these reasons, they have been the international focus of observational and theoretical efforts, as well as of laboratory measurements of the nuclear reactions that occur in (and sometimes drive) the explosions. Nova explosions are accretion-driven phenomena, caused by the transfer of mass from one star to a compact white dwarf companion. The mass transfer and subsequent rise in temperature and pressure can initiate a runaway thermonuclear explosion, resulting in the synthesis of heavy elements (to mass ~ 40) and subsequent ejection into space. These catastrophic stellar events are characterized by extremely high temperatures and densities - greater than 108 K and 103 g/cm3, respectively. Such conditions enable (p, y) and (oc,p) reactions to rapidly (on timescales of ns - min) produce unstable nuclei on the proton-rich side of the valley of stability. Any such nuclei (which decay via ^-emission) produced with half-lives longer than, or comparable to, the mean time between nuclear reactions will become targets for subsequent nuclear processing. Sequences of (p, y) and (cc,p) reactions on proton-rich radioactive nuclei therefore occur during these explosions [1,2] and

149

150

produce abundances which are very different than those from the hydrogen burning occurring in non-explosive environments. Indeed, observations of nova outbursts have revealed an elemental composition that differs markedly from solar [3,4]. Recent theoretical studies indicate that these differences are caused by the combination of convection with explosive hydrogen burning in the degenerate layer on the surface of a white dwarf star [5]. This results in a unique nucleosynthesis that is rich in odd numbered nuclei such as l3C, l5N, and l7 0 which are difficult to form in other astrophysical environments. Some of the radioactive nuclei (those with lifetimes longer than 100 s) synthesized in explosions may be carried by convection to the top of the envelope before they decay (and make a small contribution towards powering the expansion [6]). Observations of the y-ray lines (especially the 511-keV emission of l8F) resulting from such radioactive decays in the envelope may provide stringent tests of nova models [7,8,9]. These y-ray emissions depend sensitively on the amount of radionuclides synthesized by nuclear reactions in the explosion, which in turn depends on the rates of nuclear reactions driving the explosion [10,11]. Recently, it was shown that changes in the reaction rates used within a nova simulation had significant effects on both the production of individual isotopes (which can change by orders of magnitude in some cases) and on the peak luminosity and mass of ejected material [12]. Another study determined that the amount of observable 1SF surviving the nova thermonuclear runaway and transported into the envelope is severely constrained by the rates of nuclear reactions which destroy l8F [9]. Current nova models also have difficulty reproducing global observables. For example, predictions of the mass of ejected material are in some cases a factor of 10 smaller than observations [13], and improved reaction rates will help quantify this problem. Other accretion-driven phenomena important in astrophysics include X-ray bursts and X-ray pulsars. These can occur when material is accreted onto the surface of a neutron star, where temperatures and densities can reach over 109 K. and 106 g/cm3, respectively [14,15], The ensuing explosive hydrogen burning can synthesize isotopes with masses up to masses 80 - 100 or beyond [15,16,17]. Recent studies of nucleosynthesis in these violent explosions suggest that their X-ray luminosity is influenced by the nuclear reactions (most involving proton-rich radioactive isotopes) used in the model [18]. There are also other astrophysical sites [16] - e.g., the accretion disk around black holes [19] - where nuclear reactions on proton-rich radioactive isotopes may play an important role. Critical comparisons of models for any of these sites with observations require a knowledge of the rates of nuclear reactions on radioactive isotopes. Measurements of such reactions have been, until recently, impossible due to the lack of intense radioactive nuclear beams. Models therefore employ reaction rate estimates based on systematic properties of nuclear states, on information from analogue nuclei, on partial resonance information from stable beam transfer reaction studies, and on statistical model calculations. For rates dominated by resonances, these estimates can be incorrect by orders of magnitude [2], and therefore the model

151

predictions of isotope synthesis and energy generation are necessarily uncertain. The recent development of radioactive beams has initiated a new era in laboratory nuclear astrophysics - one in which previously unattainable cross sections of nuclear reactions of astrophysical importance can be measured and subsequently incorporated into an emerging generation of sophisticated, computationally intensive models of stellar explosions. Since numerous (p,y) and (ot,p) reactions on radioactive isotopes are thought to play an important role in stellar explosions, simulations have been used to determine which reactions are most important to measure. These simulations indicate that nuclear burning occurs through the Hot CNO cycles, including reaction sequences such as 12C(p,y),3N(p,y),40(e+ ve),4N(p,y)l50(e+ ve)'5N(p,cc)l2C and 16 0(p,y)17F(p,y)l8Ne(e+ve)18F(p,a)l50 [1,20]. Nuclei may be processed out of the Hot CNO cycles to isotopes with masses greater than 20 via reaction sequences such as 150(a,y)l9Ne(p,y)2()Na(p,y)2,Mg..., l8F(p,y)19Ne(p,y)20Na(p,y)2,Mg..., or 18 Ne(a,p)21Na(p,y)22Mg.... These can lead to hydrogen burning through the rapid proton capture process (rp-process) [1,20], involving (p, y) reactions near the proton drip line competing with e+ - decay and reaction cycles (e.g., the Ne-Na and Mg-Al cycles). At very high temperatures characteristic of X-ray bursts and X-ray pulsars (T9 ~ 1.5, where we define T9 as T(K)/10 9 K), the reaction sequence 12 C(p,y),3N(p,y)140(a,p),7F(p,y)18Ne(a,p)21Na(p,y)22Mg... is the trigger for the (cc,p)process, which transitions into the rp-process at approximately mass 40 [15, 19]. The viability of any of these sequences depends on the rates of the reactions involved, and those on proton-rich radioactive isotopes have the largest uncertainties. These sequences are important because they increase the energy generation rate compared to non-explosive burning while simultaneously altering the abundances (e.g., the ratio of nitrogen and oxygen isotopes) that are synthesized. The reactions ,4 0(a,p) 17 F, 17F(p,y)18Ne, l8F(p,y)19Ne, and 18F(p,a)150 play an important role in the sequences discussed above. For these reasons, we have studied these reactions via experiments with radioactive beams at ORNL's Holifield Radioactive Ion Beam Facility (HRIBF) to better determine their rates at temperatures corresponding to stellar explosions. 2

The Holifield Radioactive Ion Beam Facility (HRIBF)

The HRIBF [21] is the only U.S. user facility producing and accelerating beams of proton-rich radioactive heavy ions using the Isotope Separator On-Line (ISOL) technique. To date, radioactive beams of 17F, 18F, 6970As, and 66'67Ga have been produced at the HRIBF; intensities of the fluorine beams were as high as a few million per second on target. Beams of 56Ni and 7Be are under development, and tests are beginning with a uranium carbide target to produce beams of neutron-rich radioactive isotopes. Radioactive ions are produced when a high-temperature (1100 - 2200° C), thin, refractory target [22] is bombarded by a 0.5 kW light ion (p, d, 3He,

152 or 4He) beam from the K=105 Oak Ridge Isochronous Cyclotron (ORIC). For example, fibrous Hf203 targets [23] are being used to produce 17F via the 160(d,n)17F reaction at 45 MeV and to produce 18F via the 16 0(a,pn) ,7 F reaction at 85 MeV. This target material was chosen because of its high surface area to volume ratio and low density (enhancing diffusion of radioactive isotopes) and high melting point to withstand intense light-ion bombardment. A number of novel target configurations for high-efficiency release of radioactive isotopes are currently in use, and others are under design [22]. Once produced, the radioactive isotopes thermally diffuse out of the hot target material and effuse through a short (10 cm) transfer tube to an ion source for ionization and extraction. The particular ion source used is chosen to maximize the extracted radioactive species. An electron beam plasma ion source [23] has been used to produce positively-charged radioactive isotopes of F, As, and Ga, and a surface ionization source has been used to produce negatively-charged radioactive isotopes of F [24]. Once produced, the radioactive ions are then chargeexchanged (if they are positive ions) in a Cs vapor cell, then undergo two stages of mass separation (with the second having AM / M ,8M 0 ). The highly en­ ergetic explosion with a typical explosion energy .Eexp ~ 1051erg is triggered by the gravitational collapse of a central core after it exceeds a certain critical density (~ 10 9 5 g/cm 3 ) or temperature (~ 10 9 7 K). The event is followed by the formation of a neutron star or a black hole. This type of supernova is essentially the same as the one formerly called Type II supernova, although the current classification is based on the mechanism of explosion rather than the spectroscopic feature. Hence some subclasses of type I supernova, namely, type Ib/Ic, are thought to belong to the collapse-driven supernova. The en­ ergy source of explosion is the gravitational binding energy liberated by the formation of a compact object, typically a neutron star. This enormous en­ ergy (ENS ~ 1053erg) which is about two orders of magnitude larger than the observed explosion energy is largely transported by neutrinos copiously emitted by a nascent neutron star. Therefore, it is reasonable to think that the neutrino transport plays an important role in producing a supernova. The main reason why the collapse-driven supernova is so fascinating is that it involves a rich variety of physics. As already mentioned above, microphysics such as the weak interaction and nuclear physics is supposed to dictate the macroscopic dynamics in a critical way. In particular, the under­ standing of the equation of state for the hot and dense nuclear matter and neutrino interactions therein are crucial. On the other hand, it has become a common sense that the collapse-drive supernova is in general not spherically symmetric. Various convections are expected to occur at different places and times, and some of them inside the supernova core might play a major role in producing an energetic explosion and some in the mantle and envelope mix up heavy elements synthesized explosively, which fact was supported observation-

181

182

ally in SN1987A. The stellar rotation is another ingredient which renders the explosion asymmetric and attracts much attention these days in connection with the rotational supernova (or better the hypernova) as a possible central engine of the gamma ray burst. It is, however, noted that SN1987A was glob­ ally asymmetric. The rotation may be important for the normal supernova as well. The interest in the magnetic field in the supernova will also be revived in the same context since the existence of highly magnetized neutron stars is now becoming realistic. All these microphysical as well as macrophysical is­ sues should be investigated to understand the collapse-driven supernova, that is definitely a major challenge. The collapse-driven supernova occupies an important position in astro­ physics. One of the reasons is that it is believed to be a major contributor for the chemical evolution of universe. The supernova distributes heavy el­ ements which are synthesized not only in the hydrostatic phase prior to the explosion but also during the explosion itself, thus increasing the metalicity of galaxies. The collapse-driven supernova is also expected to be a promising r-process site, since the matter is neutron rich due to the electron capture. As explained shortly, a strong shock wave is generated in the supernova, pass­ ing through the whole progenitor. It is, therefore, expected that high energy cosmic rays are generated around the supernova. As easily understood from the fact that the supernova is associated with the formation of a compact object, the supernova is a relativistic event. It is worth remembering that the supernova is a possible source of the gravitational radiation as massive stars are in general a rapid rotator. The expectation that the gamma ray burst is somehow associated with the collapse-driven supernova has been rising these days. The further observations of the gamma ray burst might turn out to reveal the new aspect of the supernova. Here I will briefly outline the temporal evolution of the collapse-driven supernova. It commences with the gravitational collapse of a white dwarf like core due to the reduction of pressure either by the electron capture or by the photodisintegration of irons. The collapse does not halt until the central density exceeds the nuclear saturation density and the matter recovers enough pressure to prevent further collapse. During this collapsing phase, neutrinos emitted by the electron capture are trapped inside the core after the density becomes larger than about 10 12 g/cm 3 and the matter becomes opaque even for the neutrino. As a result, the electron capture is ceased because of the fermi blocking for emitted neutrinos. At this point, the electron fraction is somewhat larger than 0.3. Since the entropy of the core remains low ~ kB and the electron fraction is this large, nuclei survive just up to the saturation density. Incidentally, the collapsing core is divided into two part; the inner

183

part shrinking subsonically and self-similarly referred to as the inner core, and the outer part falling supersonically called the outer core. Their masses are one of the important factor for the successful explosion. When the pressure gets sufficiently large due to the nuclear force, the inner core bounces and the shock wave is generated at the boundary of the inner and outer cores. The inner core serves as a piston to launch the shock wave out through the outer core. When the shock wave somehow reaches the surface of the core, we regard the explosion as successful, since the loosely bound outer envelope is hardly an obstacle for the shock wave. As mentioned above, heavy elements are synthesized explosively as the shock wave passes through the envelope, and when it breaks out the progenitor's surface, the familiar optical display of the supernova emerges. In the above sequence of events, what is most uncertain is how the shock wave manages to propagate the outer core up to the surface of the core. In fact, almost all elaborate simulations done so far showed that the initial shock energy is not large enough to push the shock beyond the core surface and the shock wave stalls somewhere inside the outer core, turning to be an accretion shock there. This failure of the so-called prompt explosion is understood by the following simple energetic consideration. According to numerical simula­ tions, the initial shock energy is about 5 x 1051erg. As the shock propagates through the outer core, it loses its energy due to the endothermic photodisintegration of nuclei which consumes about lO 5 1 erg/M 0 and to the neutrino emission. Hence, if the mass of the outer core is larger than about 0.5M@, then the shock will be stagnated by the dissociation of nuclei alone. The inner core mass is roughly equal to the Chandrasekhar mass corresponding to the electron fraction diminished by the electron captures (note that the Chandrasekhar mass is about 0.7M© for the electron fraction ~ 0.35, since the Chandrasekhar mass is proportional to the square of the electron frac­ tion). Thus, it is easy to understand that the prompt explosion is difficult for a massive iron core. It is, however, noted that the success of the prompt explosion is dependent not only on the initial core mass but also on the inner core mass which is determined by the electron fraction that in turn crucially relies on the'rates of electron captures on protons inside and outside nuclei. The initial shock energy is also an uncertain factor which is affected by the electron fraction again and by the nuclear equation of state as well. It is, therefore, impossible at present to rule out the prompt explosion. Despite these uncertainties, the theoretical research of the collapse-driven supernova has been done mainly in the context of the so-called delayed ex­ plosion, in which a stalled shock wave is reenergized by neutrinos copiously emitted out of the proto neutron star. As mentioned already, the gravita-

184

tional energy of the proto neutron star, E^s ~ 10 53 erg, that is liberated by the collapse of an iron core, is mainly transported by these neutrinos. This is so large an amount that it is not unexpected that many researchers have devoted themselves to this scenario. Again, however, the subtlety comes from microphysics. Since neutrinos interact with matter only through the weak interaction, it seemed that the deposition of their enormous amount of energy to matter is not large enough to revive the shock wave. The quantitative calculation of the efficiency of energy deposition, however, requires accurate estimates of the weak interaction rates as well as the sophisticated numerical treatment of neutrino transport. This has been the topics for the last couple of years. The multi-dimensional aspects of the dynamics of supernova has also been studied extensively. Among other things, the convection in various regions has attracted much attention. Since the microphysics in the hot and dense matter looks so complicated that it is justifiably expected that some robust hydrodynamical effect is responsible for the supernova explosion. Unfortunately, this has not been vindicated as yet. It is, however, no doubt that more effort will be dedicated to the multi-dimensional study for the next years to come, provided the possible link with the gamma ray burst. In the following sections, I will focus on some particular issues as briefly mentioned above and emphasized in the table on next page. The importance of these issues is, hopefully, well understood already from the above short introduction.

2

ID

In this section, I will discuss some issues which have been studied mainly by ID simulations. 2.1

Uncertainties in the prompt explosion

The difficulty in the prompt explosion lies mainly in the insufficient energy of the shock wave at the generation and the ensuing large energy loss. As men­ tioned in the introduction, the latter factor is determined by the proportion of the masses of the inner and outer cores as well as the total core mass. 1 The inner core mass is roughly equal to the Chandrasekhar mass corresponding to the electron fraction. The electron fraction is decreased due to the electron captures on free protons and nuclei, the latter of which is more important in the lower density regime (

Figure 1: Mechanism of occurrence of kaon r-. „ „ „ . . . „„ s . Figure 2: Maxwell construction: c reprecondensation. Sobd bne represents single par. ... , j .. T^ , r i i iC sents critical density, Jt/qual pressure region t i d e energy for kaons e and dotted-bne, c o n t i n u e f r o m d e n s i t y "N" to density "K". electron chemical potential /i e .

on chiral symmetry to treat fluctuations around the condensate 5 ' 6 . We study here the properties of kaon condensed matter based on the framework and discuss the behavior of PNS, especially, the possibility of the delayed collapse. 2

Numerical Results and Discussions

With thermodynamic potential in the reference5,6, we, hereafter, use the heavybaryon limit for nucleons5. We show the phase diagram, the EOS and then discuss the properties of a PNS where thermal and neutrino-trapping effects are very important. 2.1

Kaon Condensed Matter: Phase Diagram and EOS

First, the phase diagram is shown in Fig.3. In the neutrino-trapping case, we set Yie = Ye + Y„e = 0.4 where Ye{Yv 0 in the neutrino-trapping case while n„t = 0 in the neutrino-free case, which means that once neutrinos are trapped, kaons should wait to condense until its energy further decreases. Both effects stiffen the EOS (see Fig.4) and are remarkable around the critical

196 100

-

90 -

350

IIII ;

80 300 70 60

>

50

■f

^>W v^^O**^ from Mainz experiment V ^ >O , ^ v . i 111,= leV \\ O O S L rrY = 0. leV \\ ^ ^ ^ ^ m>0.01eV

o O

-30 -

•-;>... -:;;^

\ \ Cross section v \ with summation Yv\ m = leV \ \ rrY=0.1eV \V\nX=0.01eV

~~-- ~~- "~~ v \ \ [

-40 -

Cross section" ~ - - ~ N S \ . without summationKsN>^^

-

/

\S^>^

m = leV

^^0~----__^3^!^ ~mz===://

n\=0.1eV rn^,- O.Olev

-50 -

/

S

/mv=0eV i

-10

log 10 p v (eV) _i

i

i

i_

-5

-10

log10 X/27t (cm) Figure 1. Cross sections are plotted as a function of incident neutrino momentum pv and the wave length X/2n assuming that the neutrino number density is 102 c m - 3 . The dashed line denotes the cross section without the summation over the incident waves, the solid line denotes the cross section with the summation, and the evaluated cross sections from the data are plotted as the thicker lines. Since the cross section depends on the assumed electron neutrino mass, the cross sections for the mass of 0 eV, 1 eV, 0.1 eV and 0.01 eV are shown in the plot respectively.

be consistent with the position of the induced peak in our fitting using the

220

following massive Fermi distribution; . nv(mve,(j,eff)

A-K{2mvekTvf/2

. =

73 h

f°° /

Jo

Va dx

-.

l + exp \x-

^,

(7)

!$£)

where n„ = 102 c m - 3 and Tv = 5.8 x i O - 1 0 eV are used following the scenario. Solving the equation (7) numerically, the electron neutrino mass is plotted with respect to the absolute value of the effective chemical potential in figure 2. As seen from the figure, as long as the effective chemical potential is positive, there is a room which does not conflict with the position of the induced peak in our fitting. 4

Summary

We have attempted to explain the anomaly of the negative mass square of the electron neutrino observed in the tritium-/? experiments by the reaction with the relic neutrino. Introducing a 5-function as the induced peak by the relic neutrino into the fitting parameterization of the Mainz Group, we have found that our parameterization resulted better reduced-x 2 in the physical region of mu2 > 0 eV2.(See reference11.) Although the best value of the mass square was consistent with zero as the result of the fitting, if the finite energy resolution of the detector is taken into account, there is still a possibility for the electron neutrino to have the finite mass. We have evaluated the cross section from the event rate found in the induced peak based on the standard number density of the relic neutrinos, and also derived the cross section for the process ve + n —* p + e~ based on the weak interaction. Although the theoretically derived cross section is much smaller than the observed cross section, if the neutrino sea could be in a coherent state and the temperature could be 5.8 x 1 0 - 1 0 eV, the cross sec­ tion would be increased by the summation of amplitude of incident neutrino waves even with the standard neutrino number density. However, this tem­ perature contradicts with assumed number density, as far as neutrinos follow the massless Fermi distribution in the radiation dominated epoch. Therefore we attempted to introduce a scenario that an effective neutrino degeneration due to the spatially inhomogeneous distribution, had occurred later than the radiation dominated epoch, where the neutrino must follow the massive Fermi distribution as far as neutrino mass is finite. Based on this scenario we found that there is a allowed region where the electron neutrino mass is consistent with the position of the induced peak in our fitting, as far as the effective chemical potential is positive.

221

>

= Heff °

10 10

-3 b-

10

-4 b-

10

. 1 ilillll

J. 1J JU.nl

L.U.UI.ul

i|l

i,

I

i i i iiml

i i i inul

i

I

IOWIO-SO-'IO-SO-'IO^IO-'IO-2

iMul (cV) Figure 2. The electron neutrino mass is plotted with respect to the absolute value of the effective chemical potential solving the equation (7) numerically.

5

Discussion

The most obscure point is how such a coherent state of neutrino sea could be produced despite of the fact that neutrinos are fermions. As a known phenom­ ena which causes a coherency over a long distance even with fermions is the superconductivity in solid state materials with extremely low temperature. It

222

is understood that Cooper pairs of electrons form via phonon exchanges and they behave as bosons. In analogy to this phenomena, if Cooper pairs of the neutrinos could form in universe, they could be in a state of Bose-Einstein condensates which would cause a coherent state over wide ranges. In order to produce Cooper pairs, in general two critical conditions are necessary; the first one is that Fermi surface must form, and the second one is that there must be an attractive force between fermions. It is known that even if the attractive force is quite weak, as far as the first condition is satisfied, Cooper pairs would be produced. Therefore it is very exciting to study whether there are attractive channels between neutrinos. In particular, if attractive channels via lepton or quark exchanges could exist, spatially inhomogeneous neutrino condensates might be expected. References 1. A.A.Penzias and R.Wilson, Astrophys. J. 142, 419 (1965). 2. Edward W.Kolb,Michael S.Turner, The Early Universe, (Addison Wesley, 1990). 3. Ch.Weinheimer et al., Phys. Lett. B 300, 210 (1993). 4. A.I.Belsev et al., Phys. Lett. B 350, 381 (1992). 5. E.Holzschuh,M.Fritschi and W.Kiindig, Phys. Lett. B 287, 381 (1992). 6. H.Kawakami et al., Phys. Lett. B 256, 105 (1991). 7. R.G.H.Robertson et al., Phys. Rev. Lett. 67, 957 (1991). 8. Giani Simone, hep-ph/9712265. 9. Rabindra N.Mohapatra,Shmuel Nussinov, Phys. Lett. B 395, 63 (1997). 10. Francesco Vissani, Phys. Lett. B 413, 101 (1997). 11. O.Jinnouchi and K.Homma Phys. Lett. B 435, 381 (1998). 12. Steven Weinberg, Phys. Rev. 128, 1457 (1962). 13. J.M.Irvine and R.Humphreys, J.Phys.G 9, 847 (1983). 14. A.Picard et al., Nucl. Instrum. Methods B63, 345 (1992). 15. P.F.Smith, IL NUOVO CIMENTO 83A, 263 (1984). 16. P.F.Smith, RAL-91-017. 17. M.Kawasaki K.Kohri and K.Sato, Astrophys. J. 490, 72 (1997).

N U C L E O S Y N T H E S I S IN H Y P E R N O V A E K. N O M O T O 1 ' 2 , K. M A E D A 1 , T . N A K A M U R A 1 , K. I W A M O T O 3 , P.A. M A Z Z A L I 2 ' 4 , I.J. D A N Z I G E R 4 , F . P A T A T 5 1 Department of Astronomy, School of Science, University of Tokyo, Tokyo, Japan 2 Research Center for the Early Universe, School of Science, University of Tokyo, Tokyo, Japan 3 Department of Physics, College of Science and Technology, Nihon University, Tokyo, Japan 4 Osservatorio Astronomico di Trieste, via G. B. Tiepolo, Trieste, Italy ^European Southern Observatory, Garching, Germany We discuss the properties of the hyper-energetic Type Ic supernovae (SNe Ic) 1998bw and 1997ef. SNe Ic 1998bw and 1997ef are characterized by their large luminosity and the very broad spectral features. Their observed properties can be explained if they are very energetic SN explosions with the kinetic energy of .EK > 1 x 10 5 2 erg, originating probably from the core collapse of the bare C + O cores of massive stars (~ 30 — 4 0 M Q ) . At late times, both the light curves and the spectra suggest that the explosions may have been asymmetric; this may help us understand the claimed connection with GRB's. Because these kinetic energies of explosion are much larger than in normal core-collapse SNe, we call objects like these SNe "hypernovae". The mass of 5 6 Ni in SN 1998bw is estimated to be as large as 0.5 - 0.7 M Q from both the maximum brightness and late time emission spectra, which suggests that the asymmetry may not be extreme.

1

Introduction

Recently, there have been an increasing number of candidates for the gammaray burst (GRB)/supernova (SN) connection, which include GRB980425/SN Ic 1998bw, GRB971115/SN Ic 1997ef, GRB970514/SN Iln 1997cy, GRB980910/SN Iln 1999E, and GRB991002/SN Iln. The first example of such a candidate was provided by SN 1998bw. SN 1998bw was discovered in the error box of GRB980425 (Kulkarni et al. 1998), only 0.9 days after the date of the gammaray burst and was very possibly linked to it (Galama et al. 1998). Early spectra of SN 1998bw were rather blue and featureless, showing some similarities with the spectra of Type Ic SNe (SNe Ic), but with one major dif­ ference (Fig. 1): the absorption lines were so broad in SN 1998bw that they blended together, giving rise to broad absorption trough separated by appar­ ent 'emission peaks' (Patat et al. 2000). This supernova was immediately recognized to be very powerful and bright (Fig. 2). The SN was very bright for a SN Ic: the observed peak luminosity, L ~ 1.4 x 1043 erg s _ 1 , is almost ten times higher than that of previously known

223

224 SNe Ic near maximum

SN 1998 bw, 11 May, 1=16 days

5 Dec, t=1S days

6000 8000 Rest Wavelength (A)

Figure 1: Observed spectra of Type Ic supernovae 1998bw, 1997ef, and 19941. -i—i—i—I—i—i—i—i—i—i—i—i—i—i—i—i—i—i—|—r

'Co decay

C 0 1 3 8 & S N 1998bw

50

100

150

2W

Time since the explosion ( days)

Figure 2: Absolute magnitudes of Type Ic supernovae: the ordinary SN Ic 19941, and the hypernovae SN 1998bw (Galama et al. 1998) and SN 1997ef (Iwamoto et al. 2000) together with their models. The dashed line indicates the 5 6 Co decay rate.

225

SNe Ib/Ic. Models which described the SN as the energetic explosion of a C+O core of an initially massive star could successfully fit the first 60 days of the light curve (Iwamoto et al. 1998; hereafter IMN98). The very broad spectral features and the light curve shape have led to the conclusion that SN 1998bw had an extremely large kinetic energy of explosion, EK ~ 3 x 1052 ergs (IMN98; Woosley, Eastman, & Schmidt 1999). This is more than one order of magnitude larger than the energy of typical supernovae, thus SN 1998bw was termed a "hypernova" (IMN98). "Hypernova" is a term we use to describe the events of EK > 1052 erg without specifying whether the central engine is a collapsar or magnetar or pair-instability. SN 1997ef was also noticed for its unique light curve and spectra (Figs. 1, 2). The spectral similarities between SN 1997ef and SN 1998bw suggest that SN 1997ef may also be a hypernova (Iwamoto et al. 2000). In Figure 2 the visual light curve of SNe 1998bw (Galama et al. 1998) and 1997ef (Garnavich et al. 1997) are compared with the ordinary SN Ic 19941. Despite the spectral similarity, the light curve of SN 1997ef is quite different from those of SN 1998bw and SN 19941. Since the light curves are rather diverse, even in this limited number of samples, a range of energies and/or progenitor masses of SN Ic explosions may be implied. Here we summarize the photometric and spectroscopic properties of these hypernovae and the estimated explosion energies and ejecta mass using the hydrodynamical models. We also study nucleosynthesis in hypernovae. 2

Explosion Models for Supernovae and Hypernovae

We construct hydrodynamical models of an ordinary SN Ic and a hypernova as follows. Since the light curve of SN Ic 19941 was successfully reproduced by the collapse-induced explosion of C+O stars (Nomoto et al. 1994; Iwamoto et al. 1994), we adopted C+O stars as progenitor models for SNe 1997ef and 1998bw as well. We calculate the light curves and spectra for various C+O star models with different values of E^ and M e j. These parameters can be constrained by comparing the calculated light curves, the synthetic spectra, and the photospheric velocities with the data of SNe 1998bw and 1997ef. These model parameters are summarized in Table 1, together with model C021 for SN 19941. The position of the mass cut is chosen so that the ejected mass of 56 Ni is the value required to explain the observed peak brightness of SN 1997ef and SN 1998bw by radioactive decay heating. The compact remnant in CO60 is probably a neutron star because Mcut = 1.4 M Q , while it may be a black hole in CO100 and C0138 because M c u t may well exceed the maximum mass of a stable neutron star.

226

Table 1. Parameters of the C + 0 star models model C021 CO60 CO100 C0138H C0138L

3 3.1

(M0)

Mc+o

Mej

-15 -25 - 3 0 - 35 -40 -40

2.1 6.0 10.0 13.8 13.8

0.9 4.4 7.6 10 11

Mms

66

Ni mass

Mcut

0.07 0.15 0.15 0.5 0.5

1.2 1.4 2.4 4 3

EK(

10 5 1 erg)

1 1 8 60 30

SN 19941 1997ef 1998bw 1998bw

SN 1997ef Light Curve Models

In Figure 3 we compare the calculated V light curves for models CO60 and CO100 with the observed V light curve of SN1997ef. We adopt a distance of 52.3 Mpc. The light curve of SN 1997ef has a very broad maximum, which lasts for — 25 days. The light curve tail starts only — 40 days after maximum, much later than in other SNe Ic. The light curve of SN 1997ef can be reproduced basically with various explosion models with different energies and masses. In general, the properties of the light curve are characterized by the decline rate in the tail and the peak width, Tpeak- The peak width scales approximately as r p e a k oc K

^ M ^ E ^ ' \

(1)

where K denotes the optical opacity (Arnett 1996). This is the time-scale on which photon diffusion and hydrodynamical expansion become comparable. Since the model parameters of CO100 and CO60 give similar Tpeak, the light curves of the two models look similar: both have quite a broad peak and reproduce the light curve of SN1997ef reasonably well (Figure 3). The light curve of SN 1997ef enters the tail around day 40. The subsequent slow decline implies much more efficient 7-ray trapping in the ejecta of SN 1997ef than in SN 19941. The ejecta of both CO100 and CO60 are fairly massive and are able to trap a large fraction of the 7-rays, so that the calculated light curves have slower tails compared with C021. However, the light curves for both models decline somewhat faster in the tail than the observations. In §4.1, we will suggest that such a discrepancy between the early- and late-time light curves might be an indication of asphericity in the ejecta of SN 1997ef and that it might be the case in those SNe lb as well.

227

Time (days)

Time since Ihe explosion (days)

Figure 3: Left panel: Calculated Visual light curves of CO60 and CO100 compared with that of SN 1997ef. Right panel: Evolution of the calculated photospheric velocities of CO60 and CO100 (solid lines) compared with the observed velocities of the Si II 634.7, 637.1 nm line measured in the spectra at the absorption core.

3.2

Synthetic Spectra

To strengthen the arguments in §3.1, we compare the emergent spectra for the two explosion models. Using detailed spectrum synthesis, we can distinguish between different models more clearly, because the spectrum contains much more information than a single-band light curve. Around maximum light, the spectra of SN 1997ef show just a few very broad features, and are quite different from those of ordinary SNe Ib/c, but similar to SN 1998bw. However, at later epochs the spectra develop features that are easy to identify, such as the Ca II IR triplet at ~ 8200A, the 0 I absorption at 7500 A, several Fe II features in the blue, and they look very similar to the spectrum of the ordinary SN Ic 19941. We computed synthetic spectra with a Monte Carlo spectrum synthesis code using the density structure and composition of the hydrodynamic models CO60 and CO100. The lines in the synthetic spectra computed with the ordinary SN Ic model CO60 are always much narrower than the observations. This clearly indicates a lack of material at high velocity in model CO60, and suggests that the kinetic energy of this model is much too small. Synthetic spectra obtained with the hypernova model CO100 for the same

228

4000

6000 8000 Rest Wavelength (A)

10*

Figure 4: Observed spectra of SN 1997ef (bold lines) and synthetic spectra computed using model CO100 (fully drawn lines).

3 epochs are shown in Figure 4. The spectra show much broader lines, and are in good agreement with the observations. In particular, the blending of the Fe lines in the blue, giving rise to broad absorption troughs, is well reproduced, and so is the very broad Ca-0 feature in the red. The two 'emission peaks' observed at ~ 4400 and 5200A correspond to the only two regions in the blue that are relatively line-free. 4 4-1

S N 1998bw Model Ught curves

We calculate the light curve of SN 1998bw using progenitors of different masses and explosions of different energies. We find that the model that give the best agreement to both the light curve and the spectra is that of the explosion of a 13.8M© C+O star, ejecting 10 M 0 of material with EK = 6 x 1052 erg, including 0.5M® of 56 Ni (C0138H). In Figure 5 we compare the bolometric light curves of model C0138H (solid) with the V photometry of SN 1998bw. However, C0138H has difficulties reproducing the apparently exponential decline after day 60. On the other hand, C0138L (EK = 3 X 1052; dashed) is in better agreement after day 90, although the early light curve and photospheric velocities do not fit well.

229

0

100

200 300 400 time since the GRB (days)

500

600

Figure 5: The light curves of models C0138H (J5 K = 6 x 10 5 2 erg; solid) and C0138L ( B K = 3 X 10 5 2 erg; dashed) compared with the observations of SN1998bw (Galama et al. 1998; McKenzie & Schaefer 1999). A distance modulus of fi = 32.89 mag and Av = 0.0 are adopted. The dotted line indicates the energy deposited by positrons for C0138H.

After day ~ 200 the decline of the model light curve becomes slower, and it approaches the half-life of 56 Co decay around day 400. At t > 400 days most 7-rays escape from the ejecta, while positrons emitted from the 56 Co decay are mostly trapped and their energies are thermalized. Therefore, positron deposition determines the light curve at t > 400 days (dotted line in Fig. 5). If the observed tail should follow the positron-powered light curve, the 56 Co mass could be determined directly. The comparison between SN 1998bw and the model light curve of C0138H (which fits better at early phases) and C0138L (which is better for late phases) in Figure 5 suggests a departure from spherical symmetry. 4-2

Early Time Spectra

We have used model C0138H as a basis to compute synthetic spectra for the near-maximum phase of SN 1998bw. In Figure 6 we show the synthetic spectra obtained for 3 epochs. The synthetic spectra clearly improve over those of IMN98. In particular, those absorptions not due to broad blends, i.e. the Si II feature near 6000A, and the O I+Ca II feature between 7000 and 8000A are now much broader, in significantly better agreement with the data. Our fits at least demonstrate that a large E-& is necessary, and that a Type Ic SN O-dominated composition yields quite a reasonable reproduction of the observations.

230

4000

6000 Rest. Wavelength (A)

8000

Figure 6: Observed spectra of SN1998bw (full lines) and synthetic spectra calculated using model C0138H (EK = 6 x 10 B2 erg; dashed lines).

4-3

Late Time Evolution

In Fig. 7, a nebular spectrum of SN 1998bw on 12 Sept 1998 (rest frame epoch 139 days) is compared to synthetic spectra obtained with a NLTE nebular model based on the deposition of gamma-rays from 56 Co decay in a nebula of uniform density. Two models were computed. In one model (dotted line) we tried to reproduce the broad Fell] lines near 5300A. The 56 Ni mass is 0.65 M©, and the outer nebular velocity is 11,000km s" 1 , and the O mass is 3.5M Q . The average electron density in the nebula is log ne = 7.47 c m - 3 . In the other model (dashed line), we tried to reproduce only the narrow [01] 6300A emission line. The 56 Ni mass estimated from the broad-line fit is comparable to the value obtained from the light curve calculations. At early times, the fast-expanding lobes were much brighter than the rest, and so we observed the broad-lined spectra and the bright light curve. The fastmoving regions rapidly became thin, though, and soon emission lines appeared. Initially those were broad, dominated by the hyper-energetic lobes. Later, though, the 7-rays from the fast-moving 56 Co could escape that region more and more easily, and a significant fraction of them could penetrate down into the low-velocity region and excite the O and Mg there. Both the need for a high density region and the velocity inversion as well

231 15 SN 1998bw day 139

10

5

4000

6000

8000

10'

Figure 7: A nebular spectrum of SN 1998bw on 12 Sept 1998 (rest frame epoch 139 days) is compared to synthetic spectra obtained with a NLTE nebular model based on the deposition of gamma-rays from 6 6 C o decay in a nebula of uniform density (see text for details).

as polarization measurements (Patat et al. 2000) might indicate that the ex­ plosion is aspherical. If the outburst in SN 1998bw took the form of a prolate spheroid, for example, the explosive shock along the long axis was probably strong, ejecting material with large velocities and producing abundant 56 Ni. In directions away from the long axis, on the other hand, oxygen is not much burned and the density is high enough for 7-rays to be trapped even at ad­ vanced phases, thus giving rise to the slowly declining tail. We note that our estimate of the 56 Ni mass of ~ 0.6M© from the neb­ ula spectra in Figure 7 does not much depend on the asphericity. This is in good agreement with the spherical models C0138. Since Hoflich et al. (1999) suggested that the 56 Ni mass can be as small as 0.2 M 0 if aspherical effects are large, our results suggest that the aspherical effects might be modest in SN 1998bw.' 5

Possible Evolutionary Scenarios to Hypernovae

Here we classify possible evolutionary paths leading to C+O star progenitors. In particular, we explore the paths to the progenitors that have rapidly rotating cores with a special emphasis, because the explosion energy of hypernovae may be extracted from rapidly rotating black holes (Blandford & Znajek 1977). (1) Case of a single star: If the star is as massive as M m s > 40 M Q , it could lose its H and He envelopes in a strong stellar wind (e.g., Schaller et al.

232

1992). This would be a Wolf-Rayet star. (2) Case of a close binary system: Suppose we have a close binary system with a large mass ratio. In this case, the mass transfer from star 1 to star 2 inevitably takes place in a non-conservative way, and the system experiences a common envelope phase where star 2 is spiraling into the envelope of star 1. If the spiral-in releases enough energy to remove the common envelope, we are left with a bare He star (star 1) and a main-sequence star (star 2), with a reduced separation. If the orbital energy is too small to eject the common envelope, the two stars merge to form a single star (e.g., van den Heuvel 1994). (2-1) For the non-merging case, possible channels from the He stars to the C+O stars are as follows (Nomoto, Iwamoto, & Suzuki 1995). (a) Small-mass He stars tend to have large radii, so that they can fill their Roche lobes more easily and lose most of their He envelope via Roche lobe overflow. (b) On the other hand, larger-mass He stars have radii too small to fill their Roche lobes. However, such stars have large enough luminosities to drive strong winds to remove most of the He layer (e.g., Woosley, Langer, & Weaver 1995). Such a mass-losing He star would corresponds to a Wolf-Rayet star. Thus, from the non-merging scenario, we expect two different kinds of SNe Ic, fast and slow, depending on the mass of the progenitor. SNe Ic from smaller mass progenitors (channel 2-1-a) show faster light-curve and spectral evolu­ tions, because the ejecta become more quickly transparent to both gamma-ray and optical photons. The slow SNe Ic originate from the Wolf-Rayet progeni­ tors (channels 1 and 2-1-b). The presence of both slow and fast SNe Ib/Ic has been noted by Clocchiatti & Wheeler (1997). (2-2) For the merging case, the merged star has a large angular momentum, so that its collapsing core must be rotating rapidly. This would lead to the formation of a rapidly rotating black hole from which possibly a hyper-energetic jet could emerge. If the merging process is slow enough to eject the H/He envelope, the star would become a rapidly rotating C+O star. Such stars are the candidates for the progenitors of Type Ic hypernovae like SNe 1997ef and 1998bw. If .a significant amount of H-rich envelope remains after merging, the rapidly rotating core would lead to a hypernova of Type Iln possibly like SN 1997cy (or Type lb).

233

6 6.1

Nucleosynthesis in Hypernovae Explosive Nucleosynthesis

Since hypernovae explode with much higher explosion energies than usual su­ pernovae, explosive nucleosynthesis could have some special features. Also hypernovae have shown some aspherical signatures. We calculate explosive nucleosynthesis in hypernovae in the same way as has been done for normal supernovae; we use a detailed nuclear reaction network including 211 isotopes up to 71 Ge (Thielemann, Nomoto, & Hashimoto 1996; Hix & Thielemann 1996; Nakamura et al. 1999a) (Figure 8: top). Nucleosynthesis in normal supernovae (EK = 1 X 10 51 erg) is also shown in Figure 8 (bottom) for comparison. A similar comparison is made for CO60 and CO100. The total amount of nucleosynthesis products are summarized in Table 2. From this figure, we can see the following characteristics of nucleosynthesis with the very large explosion energy. (1) The complete Si-burning region is extended to the outer, lower density region. Whether this region is ejected or not depends on the mass cut. The large amount of 56 Ni observed in hypernovae (e.g., ~ 0.5M Q for SN1998bw and O.15M0 for SN1997ef) implies that the mass cut is rather deep, so that the elements synthesized in this region such as 59 Cu, 63 Zn, and 64 Ge (which decay into 59 Co, 63 Cu, and 64 Zn, respectively) are likely to be ejected more abundantly. In the complete Si-burning region of the hypernova, elements produced by a-rich freezeout are enhanced because nucleosynthesis proceeds under lower densities than in usual supernovae. Figure 8 clearly shows a trend that a larger amount of 4He is left in more energetic explosion. Hence, elements synthesized through a-captures such as 4 0 Ca (stable), 4 4 Ti and 48 Cr (decaying into 4 4 Ca and 4 8 Ti, respectively) become more abundant. (2) The more energetic explosion produces a broader incomplete Si-burning region. The elements produced mainly in this region such as 52 Fe, 55 Co, and 51 Mn (decaying into 52 Cr, 55 Mn, and 51 V, respectively) are synthesized more abundantly with the larger explosion energy. (3) Oxygen burning takes place in more extended, lower density region for the larger explosion energy, so that the abundances of elements like 0 , C, Al are smaller. On the other hand, a larger amount of ash products such as Si, S, Ar are synthesized by oxygen burning. Figure 9 shows the abundances of stable isotopes relative to the solar val­ ues for 3 x 1052 erg and 1 x 10 51 erg. The progenitor is the 16M Q He star and products from H-rich envelope are not included. The isotopic ratios relative to 16 0 with respect to the solar values are shown. As a whole, intermediate mass nuclei and heavy nuclei are more abundant for the more energetic explosion,

234

.001

.0001

.001

.0001

4.5 M,/Ma

5

5.5

6

6.5

Figure 8: The isotopic composition of ejecta of the hypernova (EK = 3 X 10 62 erg; top) and the normal supernova (En = 1 x 10 5 1 erg; bottom) for a 16M© He star. Only the dominant species are plotted. The explosive nucleosynthesis is calculated using a detailed nuclear reaction network including a total of 211 isotopes up to n G e .

235 Eexp = 3 0 f o e 1

I

10

o

I

I

I

I

I

I

I

i

i

i i i

i

20

I < i i ' l ' ' ' ' I

i

i

i

30

i i

i

i

i i i

40 m a s s number

50

60

0

Figure 9: Abundances of stable isotopes relative to the solar values for 3 xlO 5 2 ergs and 1 x l O 5 1 ergs. The progenitor is a 16M© He star (H-rich envelope is not included).

236

Table 2. Yields of hypernova and supernova models (M 0 ) model CO60 CO100 C0138H model CO60 CO100 C0138H

c

0.082 0.58 0.11 44Ti 2.1X10 - 4 4.5X10 - 5 2.2xl0~ 4

0

Mg

Si

S

Ca

Ti

Fe

Ni

3.0

0.10 0.42 0.95

0.037 0.19 0.52

0.006 0.025 0.088

0.0003 0.0003 0.0011

0.16 0.19 0.50

0.017 0.021

56Ni

0.24 0.22 0.29 57Nj

0.15 0.15 0.50

5.7xl0~ 3 5.7xl0"3 1.5xl0~ 2

5.6 6.6

0.028

except for the elements being consumed in oxygen burning like O, C, Al. Es­ pecially, the amounts of 4 4 Ca and 4 8 Ti are increased significantly because of the enhanced a-rich freezeout. 6.2

The Mass of Ejected

56

Ni

For the study of the chemical evolution of galaxies, it is important to know the mass of 56 Ni, M( 5 6 Ni), synthesized in core-collapse supernovae as a function of the main-sequence mass M m s of the progenitor star (e.g., Nakamura et al. 1999a). From our analysis of SNe 1998bw and 1997ef, we can add new points on this diagram. Figure 10 shows M( 56 Ni) against M m s obtained from fitting the optical light curves of SNe 1987A, 1993J, and 19941 (e.g., Shigeyama & Nomoto 1990; Nomoto et al. 1993, 1994; Shigeyama et al. 1994; Iwamoto et al. 1994; Woosley et al. 1994; Young, Baron, & Branch 1995). The amount of 56 Ni appears to increase with increasing M m s of the progenitor, except for SN II 1997D (Turatto et al. 1998). This trend might be explained as follows. Stars with M m s < 25 M Q form a neutron star, producing ~ 0.08 ± 0.03 M 0 56 Ni as in SN lib 1993J, SN Ic 19941, and SN 1987A (although SN 1987A may be a borderline case between neutron star and black hole formation). Stars with M m s > 25 M© form a black hole (e.g., Ergma & van den Heuvel 1998); whether they become hypernovae or ordinary SNe may depend on the angular momentum in the collapsing core. For SN 1997D, because of the large gravitational potential, the explosion energy was so small that most of 56 Ni fell back onto a compact star remnant; the fall-back might cause the collapse of the neutron star into a black hole. The core of SN II 1997D might not have a large angular momentum, because the

237 n—i—i—|—!—r"

I ' ' ' ' I

I8DWI 98bw(lc)

97ef(lc) 87A(llp)

93J(llb)

94l(lc)

97D(llp)

10

20

25 30 Main Sequence Mass (M 0 )

35

45

Figure 10: Ejected 5 6 Ni mass versus the main sequence mass of the progenitors of several bright supernovae obtained from light curve models.

progenitor had a massive H-rich envelope so that the angular momentum of the core might have been transported to the envelope possibly via a magneticfield effect. Hypernovae such as SNe 1998bw, 1997ef, and 1997cy might have rapidly rotating cores owing possibly to the spiraling-in of a companion star in a binary system. The outcome certainly depends also on mass-loss rate and binarity. 7

Concluding remarks

We have calculated the light curves and spectra for various C+O star models with different values of £ K and M ej and reached several striking conclusions. We have shown that the spectra of SNe 1998bw and 1997ef are much better reproduced with the hypernova models than with the ordinary SN Ic model. Therefore, we suggest that SNe 1997ef, 1998ey, and 1998bw form a new class of hyper-energetic Type Ic supernovae, which we may call "Type Ic" hyper­ novae. SN 1998bw produced ~ 0.5 - O.7M0 of 56 Ni, as much as a SN la, while SN 1997ef produced less, only ~ O.15M0, but still more than in ordinary SNe Ic. The progenitor must have been a massive star, which possibly un­ derwent spiral-in of the companion star in a close binary system. Continuing observations and theoretical modeling of this interesting class of objects are certainly necessary.

238

This work has been supported in part by the grant-in-Aid for Scientific Re­ search (07CE2002, 12640233, 12740122) of the Ministry of Education, Science, Culture and Sports in Japan.

References 1. Arnett, W.D. 1996, Supernovae and Nucleosynthesis (Princeton Univer­ sity Press) 2. Blandford, R. D. & Znajek, R. L. 1977, MNRAS, 179, 433 3. Bloom, J.S. et al. 1999, Nature, 401, 453 4. Cappellaro, E., Turatto, M., & Mazzali, P. 1999, IAU Circ. No.7091 5. Clocchiatti, A., & Wheeler, J. C. 1997, ApJ, 491, 375 6. Danziger, I.J., et al. 1999, in The Largest Explosions Since the Big Bang: Supernovae and Gamma Ray Burst, eds. M. Livio et al. (Baltimore: STScI), 9 7. Ergma, E., & van den Heuvel, E.P.J. 1998, A&A, 331, L29 8. Galama, T.J. et al. 1998, Nature, 395, 670 9. Garnavich, P., Jha, S., Kirshner, R., & Challis, P. 1997a, IAU Circ. No.6778, 6786, 6798 10. Garnavich, P., Jha, S., & Kirshner, R. 1998, IAU Circ. No. 7066 11. Hachisu, I., Matsuda, T., Nomoto, K., & Shigeyama T. 1991, ApJ, 368, 27 12. Hix, W. R. & Thielemann, F.-K. 1996, ApJ, 460, 869 13. Hoflich, P., Wheeler, J.C., & Wang, L. 1999, ApJ, 521, 179 14. Hu, J. Y. et al. 1997, IAU Circ. No. 6783 15. Iwamoto, K., Nomoto, K., Hoflich, P., Yamaoka, H., Kumagai, S., & Shigeyama, T. 1994, ApJ, 437, L115 16. Iwamoto, K., Young, T.R., Nakasato, N., Shigeyama, T., Nomoto, K., Hachisu, I., & Saio, H. 1997, ApJ, 477, 865 17. Iwamoto, K., Mazzali, P.A., Nomoto, K., et al. 1998, Nature, 395, 672 (IMN98) 18. Iwamoto, K., Nakamura, T., Nomoto, K., Mazzali, P.A., Danziger, I.J, Garnavich, P., Kirshner, R., Jha, S., Balam, D., & Thorstensen, J., 2000, ApJ, 534, 660 19. Kulkarni, S. R. et al. 1998, Nature, 395, 663 20. MacFadyen, A.I. & Woosley, S.E. 1999, ApJ 524, 262 21. Mazzali, A. P., & Lucy, L. B. 1993, A&A, 279, 447 22. Mazzali, A. P. 2000, A&A, in press 23. McKenzie, E.H., & Schaefer, B.E. 1999, PASP, 111, 964 24. Nakamura, T., Umeda, H., Nomoto, K., Thielemann,F.-K., & Bur-

239

rows,A. 1999a, ApJ, 517, 193 25. Nomoto, K., & Hashimoto, M. 1988, Phys. Rep., 163, 13 26. Nomoto, K., Suzuki, T., Shigeyama, T., Kumagai, S., Yamaoka, H., & Saio, H. 1993, Nature, 364, 507 27. Nomoto, K., Yamaoka, H., Pols, 0 . R., van den Heuvel, E. P. J., Iwamoto, K., Kumagai, S., k Shigeyama, T. 1994, Nature, 371, 227 28. Nomoto, K., Iwamoto, K., & Suzuki, T. 1995, Phys. Rep., 256, 173 29. Patat, F. et al., 2000, ApJ, submitted 30. Schaller, G., Schaerer, D., Meynet, G., k Maeder, A. 1992, A&AS, 96, 269 31. Shigeyama, T., & Nomoto, K. 1990, ApJ, 360, 242 32. Shigeyama, T., Suzuki, T., Kumagai, S., Nomoto, K., Saio, H., Yamaoka,H. 1994, ApJ, 420, 341 33. Thielemann, F.-K., Nomoto, K., k Hashimoto, M. 1996, ApJ, 460, 408 34. Turatto, M., Mazzali, P. A., Young, T. R., Nomoto, K., Iwamoto, K., Benetti, S., Cappellaro, E., Danziger, I. J., de Mello, D. F., Phillips, M. M., SuntzefF, N. B., Clocchiatti, A., Piemonte, A., Leibundgut, B., Covarrubias, R., Maza, J., Sollerman, J., 1998, ApJ, 498, L129 35. van den Heuvel, E. P. J. 1994, in Interacting Binaries, ed. H. Nussbaumer k A.Orr (Berlin:Springer Verlag), 263 36. Wang, L., & Wheeler, J. C. 1998, ApJ, 504, L87 37. Woosley, S. E., Eastman, R. G., k Schmidt, B.P. 1999, ApJ, 516, 788 38. Woosley, S. E., Eastman, R. G., Weaver, T. A., k Pinto, P. A. 1994, ApJ, 429, 300 39. Woosley, S. E., Langer, N., k Weaver, T. A. 1995, ApJ, 448, 315 40. Young, T., Baron, E., & Branch, D, 1995, ApJ, 449, L51

Overabundance of Calcium in the young SNR. RX J0852-0462: evidence of over-production of

44

Ti

H. Tsunemi, E. Miyata, J. Hiraga Department of Earth and Space Science, Graduate School of Science, Osaka University, 1-1 Machikaneyama, Toyonaka, Osaka 560-0043 JAPAN E-mail: [email protected] B. Aschenbach Max-Planck-Institut fur extraterrestrische Physik, D-85740, Garching, Germany E-mail: [email protected] Recently, COMPTEL has detected 7-rays of 1157keV from 44 Ti in the direction of the SNR RX J0852-0462 1. Since « T i is a product of explosive nucleosynthesis and its half-life, r, is about 60yrs, RX J0852—0462 must be a young supernova remnant and radiation is dominated by the ejecta rather than by interstellar matter. We have detected an X-ray emission line at 4.1 ± 0.2 keV which is thought to come from highly ionized Ca. The emission line is so far only seen in the northwest shell region of RX J0852—0462. The X-ray spectrum can be weli fitted with that of thin hot plasma of cosmic abundances except that of Ca, which is overabundant by a factor of about 7. Assuming that most of the Ca is 44 Ca, which originates from 44 Ti by radioactive decay, we estimate that there is about 8 x 10~4MQ of Ca. Combining the amount of 44 Ca and the observed flux of the 44 Ti 7-ray line, the age of RX J0852-0462 is around lOOOyrs.

1

Introduction

that from the interior was thin thermal emission. They claimed that the shell region in SN1006 was a site of cosmic ray shock acceleration with energies of up to 10 GeV. Accordingly, the shell-like structure does not al­ ways stand for a thermally radiating SNR. There are several 7-ray line emissions expected in the nucleothensesis in the supernova explosion as shown in table-!. If the life is short, it will be observed only in the Supernova explosion. If it is longer than the av­ erage life of the supernova remnant, like 2GAl, it will be observed from the interstellar matter as a diffuse emis­ sion. Among them, 7-ray lines from 44Ti has moderate life which makes it a strong connection with a relatively young supernova remnant. In other words, young SNRs will be good place to search for 7-ray line emission from 44 Ti. Cassiopeia-A, one of the young SNRs, is the first source from which 7-ray emission lines from 44 Ti have been reported 3 . 44 Ti is expected to be produced in the explosive nucleosynthesis inside the SN which also pro­ duces 56Ni. 44 Ti decays into 44Sc thereby emitting two hard X-ray lines of 68keV and 78keV with a mean life­ time of 4 hours. 44Sc decays further to 44 Ca emitting a 7-ray line at 1157keV with a half-life of 60yrs 4 that has been detected with COMPTEL. In this way, 44 Ti is con­ verted into 44 Ca. Taking into account the short lifetime of 44 Ti, its detection is a direct evidence of the source being a young SNR. The discovery of this second celestial source of 44 Ti has been reported by Iyudin l: it is located in the di­ rection of the south east part of the Vela SNR where Aschenbach 5 discovered a new independent SNR, RX

A supernova (SN) explosion is a source of heavy ele­ ments in the Galaxy. The nucleosynthesized material inside the star is dispersed into interstellar space at rela­ tively high temperatures. In young supernova remnants (SNR), there are many X-ray emission lines from highly ionized elements, which tend to be diluted when mixing with ambient interstellar matter. There are two types of young SNRs from a morpho­ logical point of view: shell-like structures and those of center filled morphology. There are also two types of young SNRs from the spectroscopic point of view: thin thermal emission and power law type spectra. The Crab nebula, a young SNR, shows a centerfilledstructure with a power law type spectrum. There is a neutron star, the Crab pulsar, in its center, which is the energy source of the SNR. There are more SNRs showing shell-like struc­ tures with spectra due to optically thin, thermal emis­ sion, e.g. Cassiopeia-A, Tycho and Kepler's SNRs. The shell structure is the result of the interaction of the shock waves and corresponding shock-heating of the associated plasma. This produces thin thermal emission including many emission lines. We note that the SNe of Tycho and Kepler have been witnessed and recorded in histori­ cal documents whereas there is no record of Cassiopeia-A which is estimated to be born late in the 17 th century. The remnant of the SN occurred in AD1006, SN1006, is another young SNR. It shows a pronounced shell-like structure as shown in figure-1. Koyama et al. 2 re­ vealed that the spectra from the north-east and south­ west regions of the shell followed a power law type while

240

241

Figure 1: X-ray image of SN1006 obtained with ASCA GIS. It clearly shows a shell like structure. The spectrum from the shell shows power law type while that from the interior shows thin thermal emission.

J0852—0462, which has a radius of 1° as shown in figure2. Based on the X-ray and 44 Ti data, RX J0852-0462 is expected to be born several hundred years ago in a sky region, which in principle could have been observed from China, and therefore some record in a historical docu­ ment is expected if the SN was of standard brightness. But if it were sub-luminous the event might have been missed by the contemporaries 6 . So far, no record has been found, which reminds us of some similarity with Cassiopeia-A. We note that quite recently a footprint of a historical SN has been found in terms of nitrate abun­ dance in Antarctic ice cores. The associated SN would have occurred around 1320 ± 30 yrs 7 . 2

Observation

We have observed RX J0852-0462 with ASCA on De­ cember 21-24, 1998 as a TOO target. This observation was performed with the two CCD camera (SIS)8 and the two imaging gas scintillation proportional counters (GIS) 9 which are placed at the focal plane of four thin-foil Xray mirrors 10 on board the ASCA satellite11. The observation was performed in 7 pointings so that we could cover the entire remnant with the GIS whereas we missed the south bright shell region as shown in figure3. The SIS, having 4 times better energy resolution than the GIS, can however cover only a very small fraction of the source. The GIS observation shows a shell-like structure similar to that obtained with ROSAT. Its X-

ray spectrum is almost uniform over the source, consist­ ing of two components: a power law component with a photon index of «■' 2.5 and thin thermal emission of (1.5 ± 0.1) x 107 K. The X-ray spectrum from the entire source is shown in figure-4 by the white square. RX J0852-0462 overlaps with the Vela SNR which emits a thin thermal emission with a temperature of about 2 x 106 K. Because of the similar temperatures it is likely that the low temperature component seen in RX J0852—0462 is basically background emission from the Vela SNR. Therefore we can expect that the emission from RX J0852-0462 is predominantly of the power law type originating from a shell-like region. The SIS can cover only the central region of each GIS observation. Among them, the spectrum in the bright north-west shell clearly shows features that differ from those observed elsewhere as shown in figure-4. We see a clear emission line structure at 4.1 ±0.2 keV. We only see this structure in the eastern part of bright shell region as shown in figure-3. The other regions covered with the SIS as well show feature-less spectra similar to that obtained with the GIS for the entire remnant. 3

Discussion and conclusion

What is the origin of this emission line ? There are some candidates from an astrophysical point of view. If it is a characteristic X-ray line, it must be either from neutral Sc-K or a helium-like Ca-K emission line. From the cos-

242

Figure 2: Low energy m a p (0.1-2 keV) for Vela SNR obtained with ROSAT (left). High energy m a p (1.5-2 keV) for the same region (right). In the high energy map, the Vela Pulsar and Puppis-A SNR (upper right) are clearly seen. Furthermore, a circular structure in the lower left becomes evident that is a supernova remnant, RXJ0852-4622.

Figure 3: The circular FOV of the GIS are superposed on the X-ray intensity map obtained with ROSAT. There are two bright shell regions: north and south. A small square shows the SIS FOV detected Ca emission line.

Table 1: Decay chain from nucleosynthesis in supernova. Decay chain M

Ni

-► >*Co - .

"Co-

i7

22

Na-

44

m

m

26

Fe

Fe

22

"Ti—

Fe-*

M

1.1

A'e

Sc—

Co-*

AI-* ™Mg

Mean life (year) 0.31

3.8 44

Co

m

Ni

60

4.3 x 10 5

1.1 x 10 6

mic abundance point of view, it is very implausible that it comes from Sc. If it comes from shock-heated matter, it is plausible that it comes from Ca at high temperatures. If this is not the case, it can be either a red-shifted FeK line from an AGN or a cyclotron emission line from a strongly magnetized neutron star. Since the line emission region is quite extended judging from the ASCA image, neither one of the two hypotheses of a point-like source is very likely. We analyzed the data using a thin thermal model 12 in collisional ionization equilibrium (CIE) and we ob­ tained an acceptable fit. The temperature is (1.5 ±0.1) x 107 K. The metal abundances are similar to the cosmic abundances with the exception of Ca. The Ca abun­ dance, A, is 7 ± 4 times higher than the cosmic value. If we employ a thin thermal model with nonequilibrium collisional ionization (NEI)13, the temperature increases while the metal abundances do not change within the statistics. We cannot determine the abundances of the light elements from He to O, which produce emission lines outside our energy range. Whereas we can conclude that the relative metal abundances, from Ne to Fe, are con­ sistent with cosmic values with the exception of Ca. The overabundance of Ca can be a result of an over­ abundance of 44 Ti which is produced only in the SN ex­ plosive nucleosynthesis process. It is mainly produced deep in the interiQr of the star, both in a type la and type II SN. In the following we assume that the ejecta contain­ ing 44 Ti are uniformly expanding into ambient space. In the early phase of the SNR evolution, the major part of the emission does not come from the shock heating pro­

Energy (MeV) e+ 0.847 1.238 0.122 0.014 e+ 1.275 e+ 1.156 0.078 0.068 1.322 1.173 0.059 e+ 1.809

event/disintegration 0.2 1 0.7 0.88 0.88 0.9 1 0.94 1 1 1 1 1 1 0.85 1

cess but comes from accelerated and decelerating parti­ cles, which produce a power law type X-ray spectrum. In the north-west bright shell region, the ejecta might have recently collided with an interstellar cloud forming some shock-heated thermal plasma, which is the source of the Ca emission line. Assuming spherical symmetry for the plasma emitting region we find that the plasma density in this region is (1.4 ±0.1)(D/200 pc) -0 - 5 H cm3 where D is the distance to the source. The expanding ejecta happen to hit the interstellar cloud at the north-west shell. The ejecta expanding along other directions should also contain similar amounts of Ca but they are not yet shock-heated, so that just the power law type spectrum prevails with no emission lines. There is another bright shell region in the south, which we missed. Based on the emission line detected, we can estimate the total amount of Ca, Mc a , contained in RX J0852-0462 as 8 x l(r4M©(.D/200 pc) 25 (A/7). We as­ sume that all the 44Ca is in the shell region but only that fraction expanding towards the north-west shell is actually shock-heated and visible in the Ca line. The major isotope of Ca on the Earth is 40 Ca while the fraction of 44 Ca is about 2 x 10 - 2 14. If we assume that the other Ca isotopes are produced with terrestrial abundances 44 Ca, the product of 44 Ti, is heavily over­ produced, and we expect that almost all Ca detected must be 44Ca. Based on the theoretical model of Nomoto et al. 15 and Thieleman et al. 16, type II SN can produce Ca of 5 x 10~3 MQ depending on the progenitor star mass, whereas type I SN can produce 1.2 x 10~2 MQ. In both cases, the

244 .

■

-

1

,>t >

,

.

[

_

,

,

'l

jjjkkf

ASCAGIS

If

0.1

o o

ASCA SIS

™T itt

■

o U

0.01 : 0.001

\ 2 5 Energy (keV)

'

:

* % \ 1 2 Energy (keV)

10

HTI 5

Figure 4: X-ray spectrum from the bright shell obtained with the SIS (crosses). It can be well fitted with a thin thermal emission model of (1.5 ± 0.1) x 10T K. Metal abundances are well fitted with cosmic values with the exception of Ca which is overabundant by about 7 ± 4. The X-ray spectrum from the entire remnant obtained with the GIS is also shown for comparison.

mass ratio between Si and Ca is less or similar to that of the cosmic value. These models cannot explain the overabundance of Ca. Whereas these models assume the point symmetric explosion. Asymmetric explosions can produce large metal abundance anomalies, which may explain the overabundance of Ca. If we assume that most of Ca comes from 44 Ti, we can estimate the age, t, of RX J0852-0462 using the observed flux of the 1157keV 7rays. We obtain t / r = 16 + log2 ((L/3x 10" 5 photon cm" 2 s"1) (D/200PC)1'2 (A / 7) ). This value results in an upper limit of the age of 960 yrs and the lowest limit for the expansion velocity of 2900 km s _ 1 . Since there occurs shock-heating associated with deceleration of the expansion the initial expansion velocity must be higher than this value. If we assume that Ca has an isotope population similar to that of the terrestrial value, the mass of 44 Ti is reduced by a factor of 50. Then we obtain a lower limit of the age of 630 yrs and an upper limit of the expansion velocity of 4600 km s _ 1 . Model calculations of SN show that the heavy ele­ ments are produced relatively deep inside the progenitor star. In type II SN, 44 Ti is produced very close to the mass-cut point, which is the critical boundary separating the ejecta and the mass forming a central compact rem­ nant. If the SN explosion occurs in a homologous fashion, 44 Ti and Ca are concentrated in a narrow layer. We can expect that the entire Ca is shock-heated simultaneously. Based on the SN model calculations the expanding ve­ locity of the Ca dominated layer is a few thousand km s" 1 1 7 . Therefore, the age of RX J0852-0462 is likely to be closer to 900 yrs rather than 570 yrs.

Astro-E, the fifth Japanese X-ray astronomy satellite to be launched in February 2000, will carry three instru­ ments, which are a hard X-ray detector (HXD), an X-ray CCD camera (XIS) and an X-ray calorimeter (XRS). 44 Ti emits 3 lines at 68 keV, 78 keV and 1157keV, one of which has been detected from RX J0852-0462 by COMPTEL. T^e HXD will have sufficient sensitivity to detect the other two lower energy lines. The XIS will be able to map the distribution of Ca across RX J0852-0462. The X-ray calorimeter XRS will test whether or not the emis­ sion line at 4.1 keV originates from highly ionized Ca. References 1. Iyudin, A.F., Schonfelder, V., Bennett, K., Bloemen, H., Diehl, R., Hennsen, W., Lichti, G.G., van der Meulen, R.D., Ryan, j \ , & Winkler, C.Nature 396, 142 (1998) 2. Koyama, K., Petre, R., Gotthelf, E.V., Hwang, U. Matsuura, M., Ozaki, M., Holt, S.S. Nature 378, 255 (1995) 3. Iyudin, A.F., Diehl, R., Bloemen, H., Hermsen, W., Lichti, G.G., Morris, D., Ryan, J., Schonfelder, V., Steinle, H., Varendorff, M., & Winkler, C. Astron. & Astrophys. 284, LI (1994) 4. Ahmad I., et al.,Phys. Rev. Lett. 80, 2550 (1998) 5. Aschenbach B. Nature 396, 141 (1998) 6. Aschenbach, B., Iyudin, Schonfelder, V., Astron. & Astrophys. 350, 997 (1999) 7. Burgess, C. P. & Zuber, K. 1999, astro-ph/9909010 8. Yamashita A., Dotani T., Bautz M., Crew G.,

9.

10. 11. 12. 13. 14. 15. 16. 17.

Ezuka H., Gendreau K., Kotani T., Mitsuda K. et al. IEEE Trans. Nuc. Sci. 44, 847 (1997) Makishima K., Tashiro M., Ebisawa K., Ezawa H., Fukazawa Y., Gunji S., Hirayama M., Idesawa E. et al. Publication of Astron. Soc. Japan 48, 171 (1996) Serlemitsos P.J., Jalota L., Soong Y., Kunieda H., Tawara Y., Tsusaka Y., Suzuki H., Sakima Y. et al. Publication of Astron. Soc. Japan 47, 105 (1995) Tanaka Y., Inoue H., Holt S.S. Publication of As­ tron. Soc. Japan 46, L37 (1994) Mewe, R., Gronenschild, E.H.B.M., k van den Oord, G.H.J., Astron. & Astrophys. 62,197(1985) Masai K., Astrophys. & Space Sci. 98, 267 (1984) Anders, E., k Grevesse, N. Geochim. Cosmochim. Acta 53, 197 (1989) Nomoto, K., Thielemann F.-K., Yokoi, K., Astro­ phys. Jounal 286, 644 (1984) Thielemann F.-K., Nomoto K., Hashimoto, M., Astrophys. Jounal 460, 408 (1996) Shigeyama, T., Nomoto, K., k Hashimoto, M. As­ tron. & Astrophys. 196, 141 (1988)

X-RAY SPECTROSCOPY A N D CHEMICAL COMPOSITION IN THE UNIVERSE

Department

K. K O Y A M A of Physics, Graduate School off Science, Kyoto Univesity, Kita-shirakawa, Sakyo-ku, Kyoto 606-8503, Japan. E-mail: [email protected]

This paper reviews the chemical evolution of massive stars and their remnants using the ASCA results of X-ray spectroscopy. We demonstrate that the ASCA spectrum of Eta-Carinae provides direct evidence for the CNO cycle reaction at the surface layer of the massive star. Then we present distinct difference of line emission features in the X-ray spectra of young supernova remnants (SNR) of two types of supernovae, type la and type II. We show that the chemical composition of a galaxy can be estimated by aged SNR samples. Finally we discuss related topics of SNRs, in particular of particle acceleration in the SNRs.

1

Introduction

Heavy elements such as C, N, 0 , Si, S and Fe are synthesized in massive stars and are distributed to the interstellar space by stellar wind or by supernova explosion. This scenario has been drawn, mainly thorough theoretical ap­ proaches. In this talk, I will review the sequence of chemical evolution using our ASCA data. Our technique for chemical composition study is X-ray spec­ troscopy. We use the cosmic high temperature plasma, in which most of the outer shell electrons of heavy elements are removed, hence the atoms should be in the form of very simple structure, namely He-like or hydrogen-like. In this simple structure, X-ray energy and intensity for the transition of excited level to the ground state are highly predictable. Furthermore, since the density of our cosmic plasma is ver low, any perturbation on the atomic transition by other particle is generally negligible. Thus from the observed X-ray spectra, we can acculately estimate the chemical composition in the plasmas. 2

Nuclear burning and the CNO cycle

Hydrogen-burning starts at the core of stars and move on toward the outer region. In the late stage of a massive star evolution, the H-burning due to the CNO cycle arrives at the surface of a star. Figure 1 shows the CNO reactioncycle. Moving one trun around the reaction-cycle, we get 4 protons and emit one He nucleus, while total number of C+N+O is conserved. Since the cross section of the 14 N 4- P reaction is the minimum among all the reactions in the CNO cycle, many 14 Ns are accumulated, like a traffic jam in a belt-way. In

246

247

other words, over-abundance of nitrogen relative to oxygen and carbon is key evidence for the presence of the CNO cycle.

He I .12,

16

/-y

F

,3

X

1 7

n^

L4TOF-#

N

*J?C

Figure 1. Diagram of the CNO cycle

Eta Carinae is the most massive star in our Galaxy, possiblly more than 100 of solar mass. In the last phase of evolution, massive stars exhibit occa­ sional eruptions of the surface layer of the star. The eruption makes a high temperature plasma by the collision to circumstellar gas, hence emits strong X-rays. The ASCA satellite obtained the best X-ray spectrum from this star (Tsuboi et al. 1999). Figure 2 is the observed X-ray spectrum together with the best-fit plasma model. Our surprise was a bump at 0.5 keV. We have never seen such structure in the X-ray spectrum of any other astronomical objects. The bump can be well explained by an Lyman alpha emission of enhanced abundance of nitrogen. In fact, the best-fit model requires number ratio of N/O = 3, about 30 times larger than that of solar abundance ratio of 0.1. Therefore this spectrum provides direct evidence of the CNO cycle reaction at the surface layer of a massive star.

248

1

2 Energy [keV]

5

10

Figure 2. X-ray sepctrum of Eta Carinae

In the surface hydrogen burning massive stars, inner core should be accu­ mulated by the final product of nuclear fusion, irons. When the iron mass exceeds the Chandrasekhar limiting mass, the iron core collapses to a neutron star or a black hole, converting the huge gravitational energy to the exlosion energy of the outer layer (type II supernovae). In the case of Eta Carinae, this dramatic explosion may occur in very near future. It may be tomorrow or 1 million years latter. I really hope to continue my study of the chemical composition after the supernova explosion of this massive star. However I have no time to wait for a spectacular supernova or even hypernova event of the most massive star Eta Carinae, hence I have to move on SNR samples of other massive pogenitor stars.

3

T y p i n g of Supernovae Using t h e X-ray S p e c t r a

Figure 3 shows the 2 typical X-ray spectra after about one thousand years of supernova explosions (Hayashi 1998). The over-all structure are similar with each other. However, in detail, we can find clear contrast in the line position and flux. From the line energy, we can identify the line emitting elements.

249

■ fx\A,s M, ; S

-

/

/

U Ca

■f

1

\

F
1.78 MeV. A 4 cm thick 9 Be rod of 99.5 % enrichment (2.5 cm in diameter) was irradi­ ated. Neutrons were measured with four 60 cmHg BF 3 counters (MITSUBISHI

269 10 8 w 6 e

8 "

4 2

0 0

200

400

600

800

1000

Channel Number Figure 2: Moderation time distribution of neutrons in the polyethylene cube.

ND8534-60: 2.5 cm in diameter x 24.1 cm in effective length) embedded in a 30 cm polyethylene cube. T h e BF3 counters were located, two each vertically and horizontally, in a concentric ring at 7.5 cm from the b e a m axis. Neutron moderation time in the polyethylene was measured in 1 ms time range with an O R T E C 566 TAG module. Fig. 2 shows a moderation time distribution. T h e moderation time constant was 144 yus. Background neutrons t h a t arrived at BF3 counters time-independently made a minor contribution to neutron counting. Neutron detection efficiency was measured at 265 keV with a 2 4 N a O H + D2O source and at the average energy of 2.35 MeV with a 2 5 2 Cf source. T h e production of the monoenergetic neutron source and its calibration followed Ref. 9 . Details were reported elsewhere 1 5 . T h e energy dependence of the efficiency was calculated with the Monte Carlo code M C N P 2 6 . Fig. 3 shows the neutron detection efficiency as a function of neutron energy. T h e 1 kHz LC photons were detected with a B G O detector (2" in diameter x 6" in length) placed at the end of t h e beam line. Pile-up spectra were obtained as typically shown in Fig. 4. T h e average number of photons per b e a m pulse was determined from the ratio in average channel number between the pile-up spectrum and a single photon spectrum. T h e single photon s p e c t r u m was separately taken with a DC beam. T h e total number of photons were obtained from the frequnecy and the d a t a acquisition time. Independent test measurements were carried out with a DC beam, in which foreground and background neutrons were measured before and after each measurement with

270 12 10 8

MCNP Monte Carlo Cal.

NaOH + D O 2

g6 4 'Cf

2 0 0

500

1000

1500

2000

2500

E (keV) n Figure 3: Neutron detection efficiency as a function of neutron energy.

the laser off. Results of the pulsed- and DC-beam measurements agreed to each other within 5 %. Photoneutron cross sections were obtained from Nn

Nyen(En)Ntf

(1)

where Nn is the number of neutrons detected, JV7 is the number of LC photons incident on the B G O detector, Nt is the number of target nuclei per cm 2 , and £n(En) is the neutron detection efficiency. A correction factor, / , was needed for a thick target measurement.

(2) fi is the total 7-attenuation coefficient in 9 Be 2 7 and t is the length (4 cm) of the target material. Photoneutron cross sections measured with polarized and unpolarized pho­ tons are shown by open and solid circles in Fig. 5, respectively. T h e decay channel was assumed to be n + 8 B e . Horizontal error bars stand for the skewed energy spread of LC photons in F W H M . As for vertical errors bars,

271 5000

4000 w C 3000 3 O O 2000

1000

0 0

1000

2000

3000

4000

channel n u m b e r Figure 4: Pile-up spectrum of LC photons measured with a BGO detector.

only statistical uncertainties are shown. The results of polarized and unpolarized photon measurements at 1.9, 2.7 and 3.3 MeV agreed to each other within statistics. It is noted t h a t the present cross sections for the narrow 2.44 MeV state formed a tail on the high-energy side. T h e tail arose from the low-energy wing of LC photons. T h e effect of the asymmetric photoneutron emission on en(En) was inves­ tigated with help of the M C N P code. For the asymmetries with a/b = 1.2 1 6 and 1.0 2 in a-\-bsin29 reported at 2.76 and 2.95 MeV, respectively, the associ­ ated uncertainty was found to be 9 - 12 %. This is regarded as the systematic uncertainty in the present measurement. For comparison, cross sections measured with other photon sources are also shown in Fig. 5. T h e horizontal error bars of the large crosses 3 as well as the dotted b u m p 2 corresponding to the 2.44 MeV state ( r < 1 keV) represent the energy resolution of Bremsstrahlung measurements. For l / 2 + , the result of not the Bremsstrahlung 2>3 but the radioactive-isotope measurement 9 is consistent with the present result. T h e least-squares fit to the d a t a was performed with the cross section of Breit-Wigner form. fjp

T

*,,„(£, .i^j)

T\

2J + I

he '

= -^7TTy(^)

ryrn {Ej

_ ERT

+ (r/a)»

7-decay widths for E l and M l are expressed 1 7 , respectively, by

(3)

272

JO

B

E (MeV) Figure 5: Photoneutron cross sections for 9 B e (solid circles for polarized LC photons and open circles for unpolarized LC photons). Data taken with Bremsstrahlung (dotted line [2], large crosses [3]) and radioactive isotopes (open squares [9,10], solid squares [8], solid triangles [7], open triangle [5], slashed-open square [4], diamonds [6]) are also shown. The best least-squares fit is shown by the solid lines (thick solid line for sum, thin solid lines for breakdown).

273

Ty(El)

= —a(hc)-2E*B(Ei)

(4)

and T 7 (M1) = — a(2Mcay2E*B(Mi),

(5)

where a is the fine-structure constant, M is the proton mass, and B(El) and B(Ml) are reduced transition probabilities given in units of e 2 / m 2 and (ek/2Mc)2. T„ for s-wave neutrons from 1/2+ was taken 1 8 as r „ = 2^eR(E7

- ET).

(6)

This form of neutron decay width leads to single-level approximation of Rmatrix theory 19 . T„ for other states was taken to be constant. The least-squares fit was performed to 49 data points up to 4.5 MeV with the Powell method 2 1 . The 49 data points excluded the data for the 2.44 MeV state and included those of Fujishiro et al. 9 . The El and Ml parametrizations were employed for the positive- and negative-parity states ( l / 2 + , 5/2 + , 1/2 - ), respectively. A straight-line background (a = 0.38i?7 - 1.21 mb) was assumed in the high-energy region. The least-squares fit resulted in the x2 minimum with B(M1) = 0 for 1/2". Thus, the presence of 1/2" was not confirmed as was the case in the electron scattering. The best fit (x 2 = 2.7 per degree of freedom) is shown by the thick solid line in Fig. 5. The best fit values for the l / 2 + state are the following. ER= 1.748 ±0MMeV B(E1) = 0.107 ±0.007e 2 /m 2 r ~ r „ = 283±42fceV Attached are the 1 a uncertainties from the error matrix. The B(E1 J.) value for l / 2 + is approximately twice those of electron scattering (0.050 ± 0.02011, 0.054 ± 0.004 12 , while it is close to that (0.106+Q oie) deduced from R-matrix fit to Fujishiro data alone 22 . It was pointed out 2 2 that the discrepancy for l / 2 + between the (7,n) 9 and the (e,e') n results may be due to background subtraction in the electron scattering. Tt was shown that, good agreement between the two was obtained by integrating the (7,n) cross section up to 2 MeV 2 2 . We used the energy-integrated cross section 20 to deduce T 7 for the narrow resonance state at 2.44 MeV.

274 f°°

J

,„

3jr 2 ,fic *

Ni • (GSI exp.)

10° I

so '■a u

r

10

86

Kr->Ni m (EPAX2)

78

Ni

CO

\

t/3

!»

o

10 H

s-l

U 10

40

42

44

46

48

50

52

Neutron Number Figure 3: Production cross section for Ni isotope. Solid circle shows the production cross section in the fission-in-flight of U beam from GSI, and solid square the production cross section by the fragmentation of S6Kr beam calculated by EPAX2.

MUSES will be electron scattering from storaged RI beams using the DSR as an electron-nucleus collider. One can determine the electromagnetic form factor of many unstable nuclei for the first time. 3

Production of RI Beams

RI beams will be produced by either projectile fragmentation reaction or fission-in-flight reaction of U-beam. Projectile fragmentation has been already applied to produce secondary RI beams at various laboratories in the past, and the production cross sec­ tion and the kinematical behavior of fragments have been semi-quantitatively understood 2 . Fission-in-flight has been recently investigated at GSI 3>4>5, and it has been demonstrated to be quite powerful to produce very neutron-rich nuclei. As an example, the production cross section of Ni isotopes by the fragmentation and fission-in-flight reactions are compared in Fig. 3. The fission cross section are taken from GSI data 4 , and the cross section for fragmentation are calculat-

280

ed with a program EPAX2 6 . As one sees, the production cross sections of fission-in-flight are much larger (one to two order in magnitude) than those of fragmentation for producing all Ni isotopes. In the case of the fission reaction, fission fragments have much larger an­ gular and momentum spread than those by projectile fragmentation. For ex­ ample, the angular and momentum spread of fission fragments are typically about ±100mrad, and ±10% at the RIBF. This is why the separator, BigRIPS, has been designed to have large angular and momentum acceptances to keep reasonable transmission (about 8%) for fission fragments. Expected yields of Ni isotopes by fragmentation and fission were simulated using MOCADI 7 taking into account the BigRIPS acceptance. According to the simulation, the fission-in-flight method is found to give larger yields for all Ni isotopes than the projectile fragmentation, and the expected intensity of 78 Ni is about 1 Hz. Although one needs to check this by experiments, it is clear that fission-in-flight of the U-beam will play an important role especially for very neutron-rich r-process nuclei. 4

The Measurement of Half-Lives and Masses

The measurements of the half-lives of nuclei will be done by the stopping experiments. The half-lives of nuclei in the r-process region is expected to be an order of one second or less, and the production rate of such isotopes will be quite low. Isotopes, particle-identified by a usual Bp — AE — TOF method, will be implanted in, for example, a position-sensitive detector, and the observation of /3-decay at the same position in the detector as a function of time after the implantation will give the half-lives. A good position resolution is necessary to suppress background. A new detector system based on this method is under development at RIKEN 8 . Instead of using a position sensitive detector, a thin foil is used as an isotope catcher. Eleven thin foils are mounted on a rotating wheel, and one of foils is set on the beam line. In front of the foil, a position-sensitive plastic-fiber detector will be placed to mark the position of implantation on the foil. Just after implantation, the wheel rotates quickly to set the foil in front of a position-sensitive /3-ray detector for the half-life measurement, and simultaneously to place another foil on the beam line for further implantation. The detector system has five rotating wheels along the beam line, and allows the simultaneous half-life measurements of different isotopes which stop in different foils due to the range difference. This detector system is designed to determine the half-lives of nuclei, which are longer than 300 ms. The first

281

experiment using this new system at the present facility is scheduled in the beginning of year 2001. The same setup will be installed in the secondary beam line of RIBF, and the half-lives of nuclei along the r-process path will be determined. There will be several ways to determine the nuclear mass. One method em­ ployed in the first phase of RIBF will be the time-of-flight (TOF) measurement over a long flight path. The TOF measurements with good timing detectors of a few 10 ps resolution over 50 meter flight path will allow to determine the mass of the level of about a few 100 keV for A ~ 100 region, when the statistics is enough. The experimental hall is designed to provide such a long flight path with a momentum analyzing system. Another promising way to determine the nuclear mass is to use an accu­ mulator ring, such as ACR of MUSES. This method is quite powerful when the production rate is quite low. One is able to accumulate a nucleus of interest in the ACR until it decays, and to observe the circulating frequency. This method will allow very precise determination of the mass of a nucleus. As an example, let me introduce the mass measurements using a storage ring currently carried out at GSI. The experimental storage ring (ESR) has been used for this purpose. The ESR has been operated in isochronous mode. Two different techniques, namely non-destructive and destructive methods, have been employed for the precise mass measurement. The non-destructive method is to use the Schottky noise caused by the beam circulation in the storage ring 10 . In order to achieve high mass resolu­ tion, the RI beam must be cooled prior to the measurement. Therefore this method can be applied to nuclei whose half-lives are longer than a few seconds when one uses the electron cooler. It is reported that the mass resolving power of 350000 has been achieved. The other novel way (the destructive method) is to measure the flight time directly in a storage ring 1 1 . This new method has been recently applied at GSI 1 2 , and the nuclear mass, whose life time are shorter than a few ms, have been successfully measured with the mass resolution of 10~ 5 . A thin foil equipped with a MCP detector is set in the storage ring. When a circulating beam passes through the foil, secondary electrons are produced. The electrons provide the timing information, and one determines the circulating frequen­ cy. Since the time-of-flight is measured directly for each turn, no cooling is necessary. In the phase II of the RIBF project, the ACR operating in the isochronous mode will be a place to measure the masses of many unstable nuclei. The higher intensity of primary beams at RIBF will allow to determine the masses of more neutron-rich nuclei.

282

5

Conclusions

The determination of masses and half-lives of nuclei along the r-process path is one of the important subjects at the coming new-generation heavy-ion acceler­ ator facilities. Since they are so neutron rich, high-intensity heavy-ion beams are indispensable for their production. The RIKEN RIBF will be one of such facilities to study the r-process. High-energy and high-intensity primary beams and a large acceptance fragment separator, BigRIPS, allow us to select nuclei as secondary RI beam freely over a much wider range of proton and neutron numbers. The RIBF phase I will be completed in the year 2003. 1. E. Burbige et al, Rev. Mod. Phys. 29, 547 (1967). 2. J.A. Winger, B.M. Sherrill and D.J. Morrissey, Nucl. Instrum. Methods B 70, 380 (1992) 3. M. Bernas et al, Phys. Lett. B 331, 19 (1994). 4. M. Bernas, Proceedings of ENAM98 : Exotic Nuclei and Atomic mass­ es, edt by B.M. Sherrill, D.J. Morrissey and C.N. Davis, p.664, 1998, publisher 5. K.-H. Schmidt et al, GSI-Preprint-99-30, August 1999. 6. K. Suemmerer and B. Blank, GSI-Preprint-99-37, November 1999. 7. N. Iwasa et al, Nucl. Instrum. Methods B 126, 284 (1997). Nucl. Instrum. Methods B 70, 380 (1992). 8. S. Nishimura et al., A proposal for Nuclear Physics Experiment at RIKEN Ring Cyclotron (2000), 'Measurement of Properties of NeutronRich Nuclides Relevant to the Astrophysical r-process' 9. C D . Buchanan et al, Phys. Rev. D 45, 4088 (1992). 10. B.Schlitt et al, Nucl. Phys. A 626, 315c (1997). 11. H. Wollnik et al, Nucl. Phys. A 626, 327c (1997). 12. M. Hausmann et al., GSI report 1999, p.22.

CONNECTION B E T W E E N CRUCIAL NUCLEAR REACTION RATES A N D T H E M O D E L I N G OF A C C R E T I N G N E U T R O N STARS

M. H A S H I M O T O , O . K O I K E A N D R. K U R O M I Z U Department

of Physics,

Kyushu

E-mail:

University, Ropponmatsu, Fukuoka JAPAN [email protected]

810-8560,

M. F U J I M O T O Department

of Physics, Hokkaido University, Sapporo 060-0810, E-mail: [email protected]

JAPAN

K. A R A I Department

of Physics, Kumamoto University, Kumamoto 860-8555, E-mail: [email protected]

JAPAN

Effects of uncertainties of key nuclear reaction rates in the crucial nuclear processes on type I X-ray burst modeling have been reviewed/investigated. Special attention is devoted to the ignition condition of the thermonuclear flash. It is found that in the rapid proton capture process (rp-process), the ignition depends on the rates of 13 N ( p , 7 ) 1 4 0 , 1 4 0 ( a , p ) 1 7 F , and 1 5 0 ( a , 7) 1 9 Ne. On the other hand, for the pure helium flash suggested from a recent observation of X-ray bursts, we infer that the NCO-reaction plays a key role in triggering the flash.

1

Introduction

Type I X-ray bursts of low mass X-ray binaries (LMXBs) are considered to be thermonuclear runaways in accreted materials on the surface of neutron stars 1 ' 2 . Detailed evolutionary calculations have been performed taking into account the nuclear processes during the flash.3'4 In the mean while, many nuclear data have been revised and accumulated in these years, some of which may affect the modeling X-ray bursts. The nuclear process in the proton rich environments was investigated in detail by Wallace & Woosley5 who proposed the rapid proton capture process which has been initiated by the break out from the hot-CNO cycle. In a framework of the one zone model of constant pressure, explosive nucleosyntesis 6 and the rp-process 7 were investigated in detail using large nuclear networks. They showed clearly that not only the nucleosynthesis proceeds appreciably beyond 56 Ni but also appreciable amounts of the nuclear fuel of hydrogen and helium are left if the peak temperature exceeds 109 K

283

284 8.5 r—i

r

15

- 10

CO

O

bo 7.5

o - 5

_i 30

I 20

log P (Cgs) Figure 1. Structure of an accreting neutron star just before an ignition. Note that the temperature and the density are drawn against the pressure.

which corresponds to the case for the pressure of P > 3x10 dyn cm ; it was suggested that the fuel left unburnt may be responsible for the X-ray bursts at 10 minutes interval.8 Using the approximate network 7 , Fujimoto et al. 3 investigated X-ray bursts in detail with the full evolutionary calculations including the general relativity 9 for an accreting neutron star model. For example, Fig. 1 shows the temperature and density structure of an accreting neutron star after 2.63 x 104 s accretion; the total rest mass is 1.61067 M 0 and the radius 8.069 km; Then the luminosity is 29.6 L Q and the effective temperature is 3.95 x 106 K. In the present paper, we will investigate the thermonuclear flash with the use of an extended network up to " P d based on the network constructed by Koike et al. 10 which is coupled to the thermodynamical equation for the flash. On the other hand, in our stellar evolution calculation, we use the upto-date nuclear data and other physical inputs like screening factors. Previous approximate network which consists of 13 nuclei is updated changing the old nuclear data to new data. Then, it is shown that some crucial reactions govern

285 T

1

1

1

J

I

L

1

1

1

_l

l_

1

1

1

1

1

1

1

I

I

I

p

3x10* -

2xl(f

1x10" -

0 J

lxlOJ

50000

tl

1.5x10s

t (sec) Figure 2. Model calculation of successive light curves of X-ray bursts during the accretion onto a neutron star with reaction rate compilations up to 1984.

the ignition timing of the shell flash as inferred from the temperature profile of Fig. 1. 2

General feature of the shell flash model connected with the rp-process

Stellar evolutionary code for an accreting neutron star has been developed n with the inclusion of the results of the nucleosynthesis calculations 6 , 7 ; Evolu­ tion code has been coupled to an approximate network code which simulates the rp-process results calculated using a large network. 12 Then radiative zero boundary condition is imposed for Mr/M ~ 1 p =

GMt(M - Mr) / 4TTR4

2GMt

Re2

-1/2

, L = Lph

(1)

where L*h means that it is the luminosity observed far from the star, the total mass Mt, the total rest mass M, the rest mass inside the radius Mr, and

286 i

,... 30000

i

i

|


| VYY^SO) I 1 > ±WWW

A(?) = — / n

+ Ml')2,

(1)

Jo e{q) = e(q) - iF ~ (q2 - qF)/2MY, 2

(2)

r drj0(qr)VYY{r;

1

S0)j0(q'r),

(3)

where A(g) is the energy gap function, qp = {iir2 pyy-)1^ is the Fermi momen­ tum of y , and (F = h2q'F/2MY is the Fermi energy, with My being the hyperon mass. For convenience, the effective-rnass approximation, as in Eq.(2), is adopted for the single-particle energy e(q) with the effective-mass MY- The 1 5o-gap equa­ tion (1) with Eqs.(2) and (3) is solved numerically when the 1S'o-pairing interaction Vyy (r; 1 ^o), the y-fraction VY(P) and the effective-mass parameter mY = MY/MY are given. First we treat the A-case, by paying attention to the following points: (i) We choose two OBEP-type VAAOSO) which has been tested for reproducing the binding energy of AA pair in the double A hypernuclei ij^^Be,)^KB). One (abbreviated here to as ND-Soft) is a soft-core version of the Nijmegen-D hard-core potential (ND) l s and constructed so as to fit the i-matrix from the original ND potential. It is expressed as 3

VAA(lS0) = Y,{Vi + * • W } « r < r / w a

(4)

i=l

with Vf = {-21.337, -187.01,10853.3} MeV and Vfs = {0.19321,32.166,2035.1} MeV for A = {1.342,0.777,0.35} fm. The other (Ehime) 1 6 is the one from Ehime group, which is based on the framework of nonet mesons and SU(3) symmetry and has a soft repulsive core and the velocity-dependent short-range interaction. As shown in Fig.l, the ND-Soft potential has a stronger short-range repulsion and a deeper medium-range attraction as compared to the Ehime potential. The choice

307 of these two potentials is expected to cover the present uncertainties in the 1 5b A A interaction. Fig.l compares VAA^SO) with the 15'o NN interaction VNNCSO) from the Reid-Soft-Core potential l r , suggesting that the A- superfluidity is less favourable than that of protons so far investigated 18 .

100

ND-Soft(AA) Ehime (AA) RSC (NN)

V 50 (MeV)

1 i

(fm)

i

i

0

3 i

" 1

Ss

a 1

//

\

iV IV

-50

2

r

1

1

ii

ti 1\ l1 a

/

/

/

// /

1 ' / '

'

/' t

/ '

// '' /// ''' /' // /' / /

\ f '

-100 Figure 1: Comparison of 1So AA and NN potentials.

(ii) Generally, the resulting energy gap is sensitive to the effective mass. So it is important to take account of the p- and ^-dependences of m*A. Our m*A is taken from the G-matrix calculations performed for {n + A}-matter with the A-fraction 2/A and by using the ND potential. The reason to use the YN and YY interactions from ND potenial is that the ND gives a reasonable value of m A compatible with the A- hypernuclei d a t a 1 9 . The resulting m A (p,2M) is large, e.g., m*A ~ (0.80 —> 0.73) for J/A = 0.05 and ~ (0.84 —> 0.77) for J/A = 0.15, in contrast with those of protons; e.g., m A ~ (0.7 —> 0.5) for p = (1 —> 3)po 18- The larger m A suggests that the A-superfluidity is more likely than the p-superfiuidity as far as m A is concerned. (iii) As for the hyperon core model (i.e., ?/A(P))> we take the one from Pandharipande 7 , as a representative case, which increases with p, having a maximum value of about 11% at p ~ 6po and then decreasing gradually. In solving the gap equation at a given p, we use m*A(p) by combining the results of m*A(p, y A ) from (ii) and this

308

ND Soft

Figure 2: Critical temperature Tc of A-superfluidity as a function of density p, calculated for AA pairing interactions in Fig.l. po demotes the nuclear density.

UA(P)- Then the resulting energy gap versus density generates the density-region of neutron star cores where the A-superfluid exists. We show in Fig. 2, the results of critical temerature T c of A-superfluidity given by Tc ~ 0.66A(gi?) x 10 10 K with A in MeV. The superfluidity is realized when Tc exceeds the internal temperature (~ 10 8 K) of neutron stars. We see that the A-superfluidity is surely realized in the density region p ~ (2 — 3.7)p 0 for the NDSoft potential. For the Ehime potential, the aspect is similar but with higher Tc values and a wider density region p ~ (2 — 4.6)po- This comes mainly from the fact that the effect of short-range repulsion in the AA pairing interaction, growing with P A ( = VAP), is relatively small for the Ehime potential due to the weaker repulsive core in VAA(lS0) (Fig. 1). The restricted density region for the existence of A-superfluid, as above, means that for neutron stars with larger M (hence higher central density), a non-superfluid region takes part in their cores, lacking a mechanism necessary to supress "too rapid" hyperon cooling. This suggests that neutron stars compatible with "hyperon cooling" scenario should not be so massive, for example, M < 1.5M 0 , for the neutron star models from BJ-1H equation of state 2 0 . 3

S u p e r f l u i d i t y of E

and E

Now we discuss the possibility of E~ and S~ superfluids. For these cases, we treat the a 5o gap equation similarly to the A case. However, unfortunately our present knowledge of the E ~ E _ and H _ S ~ interactions is less certain than the A A one, due to the lack of experimental information, as mentioned in §1. So we are content to use, in the light of SU(3) symmetry, the 15'o YY interactions from the ND-Soft and the Ehime potentials for which the AA interaction has been tested by hypernuclear

-

i

i

i

i

|

i

(

.

r

21 id0 (K)

io 9 *

ND-Soft Ehime

m Y = i. yY = b°/o

F-G(A)

?'?a

1(f Figure 3: Critical temperature T c of E~ and S~ superfluids in comparison with that of A superfluid, by fixing parameters as mY = 1 and yy = 0.05, for three type of pairing interac­ tions (ND-soft, Ehime and F-G potentials).

data. The 1SQ E ~ E ~ and S~H~ pairing interactions corresponding to the NDSoft potential are constructed quite similarly to the AA case and are given in the form of Eq.(4), with parameters somewhat changed; Vi (7 = c, ss) —> CaV^ for i = l,2 and V^ -> C r K. (7) for i = 3 with Ca = 1.37(0.88) and Cr = 1.85(1.08) for E-E-(S-H-). Results for E~- and S~- superfluidities are shown in Fig.3 by focusing attention on a comparison with the A case for the same mY(= 1.0) and yy{= 5%) parameters. It is observed that for the Ehime potential (dashed lines) the Tc both of E~ and H~ are much larger than that of A. For the ND-Soft potential (solid lines), Tc(E~) is in a similar situation as the Ehime case (T C (E") > > TC(A)), but TC(H~) is somewhat smaller than TC(A). The latter situation, however, is changed as TC(E~) > Tc(A) when a, realistic my in medium is taken into account, since our G-matrix calculations with the ND potential suggests that m£,- and m„_ are remarkably larger than m*A; m^ > 1 and m~_ > 1 whereas m*A ~ 0.8. To summarize, we can expect that both of the E~- and H~- superfluids would be realized with the Tc comparable to or larger than that of A-superfluid. In Fig.3, we also add the results of T C (E~) and T C (S~) from another version of OBE YY potential (dotted lines) proposed very recently by Funabashi-Gifu group 2 1 , which confirms the results from the ND-Soft and Ehime potentials.

310 4

Effect of Y - m i x i n g o n s u p e r f l u i d i t y of n e u t r o n s

It has been shown that neutrons (n), a dominant component of neutron star cores (hence with high Fermi energy), become a superfluid of 3 P 2 -type (instead of x Sotype) at densities p > po due to the attractive effect of the 3 P 2 nn interaction which is most attractive at high scattering energies 14 . As the density goes higher (p > (3 — 4)po), however, this 3 P 2 n-superfiuidity becomes unlikely, because of the increasing short-mage effects in the 3 P 2 pairing interaction and the decreasing m*n. Here it is worth noting that this aspect has been obtained by neglecting the mixing of hyperons as new components. In this section, we give a comment as to the effect of y-mixing on the 3 P 2 n-superfluidity, that is, the effects coming from the decreasing yn (e.g. yn = 1 —> 0.7) with increasing yy in the dense core region (P 2 3po). The increase of the Y-population (hence the decrease of y„) has two effects. One is to lower the ep of neutrons at a given density and thereby to postpone the growth of a short-range repulsion in the pairing interaction, as compared with the case neglecting the Y-mixing. The other is to change m*n{p,yn) by a contribution from riY interaction and by a lowered qF- Clearly the former effect acts for the persistence of n-superfiuid at higher densities (p > (3 — 4)po). But the situation is defermind as a result of the combined two effects. For transparency, we simplify the problem as a comparison of Tc(n) between {n + A}-matter with yn = 0.7(J/A = 1 — yn = 0.3) and pure n-matter (yn = 1). We solve the 3 P 2 gap equation including the tensor coupling with the 3 F 2 state by adopting the OPEG sO — 1 potential 2 2 for a realistic nn interaction and by using the m*n(p,yn — 0.7) extracted from the G-matrix calculations with the ND potential. Results are shown in Fig.4. We see that the ep(n) is lowered as in Fig.4a but the m*n is made smaller as in Fig.4b, compared with those in pure nmatter. As a result of the combined effect of these, we have the Tc(n) in Fig.4c. By looking at the results at high densitites (p > 3po) where actually the Y(A)-mixing gets serious, we observe that the lowered ep(n) indeed acts for the increase of Tc(n) (a dotted line) but the inclusion of a reduced m*n compensates this advantage. It is remarked that a net effect of Y- mixing does not necessarily assist the realization of n-superfluidity at high densitites (p > 3po). If we use m* from NF (Nijmegen F) potential, the resulting Tc (n) gets still smaller, suggesting that the Y-mixing acts against the existence of n-superfluid at these high densities. 5

Concluding remarks

We have shown a strong possibility that hyperons such as A, E " and H~ in a hyperon-mixed core of neutron stars could be in a superfluid state, by paying at­ tention to the use of YY and YN interactions compatible with hypernuclear data. This result suggests that the idea of "hyperon cooling" scenario can be one of the candidates to account for an unusually low surface temperature observed for some neutron stars. The increase of hyperon components makes smaller the density (hence the Fermi energy) of neutrons, a dominant conponent, compared to the case without hyper-

311

150 100 (MeV) 50-

0.9h

0.7 _L

1

_L

-_L

2 p/o

3 (b)

n

m rn 0.8 0.7 0.6

2p,P

3

10

(k)

10°

J

2

I

9l?a

hl,.iii.iiii/lli\lii.,.,i.

3

Figure 4: (a) Comparison of neutron Fermi energy ep versus p between neutron fraction yn = 1 and 0.7 cases, (b) Neutron effective-mass m£ versus p compared between yn = 1 and 0.7 cases, calculated for Nijmegen D type (ND) and F type (NF) interactions, (c) Critical temperature Tc of neutron 3P2 superfluid for yn = 1 (a solid line) and yn = 0.7 (dashed lines). A dotted line illustrates the case where mjj for yn = 1 are used only to see the effect of lowered ep (yn = 1 —» 0.7).

312

ons, affecting the pairing attraction and the effective mass of neutrons. We have investigated this influence for a simplified system composed of n and A. It is found that the increase of A component does not necessarily work for raising the critical temperature of n-superfiuid at higher densities (p > 3po)-

Acknowledgement The authors wish to thank M. Wada for providing us with Funabashi-Gifu potential before publication. This work is indebted to a Grant-in-Aid for Scientific Research from Ministry of Education, Science, Sports and Culture (No. 11640245).

References 1. Madappa Prakash, Manju Prakash, J.M. Lattimer and C.J. Pethick, Astrophys. J. 390, L77 (1992). 2. S. Tsuruta, Phys. Rep. 292, 1 (1998). 3. D. Page, Astrophys. J. 428, 250 (1994). 4. T. Takatsuka and R. Tamagaki Proc. Int. Sym. on Physics of Hadrons and Nuclei, Tokyo, Dec. 14-17, 1998, ed. Y. Akaishi, O. Morimatsu and M. Oka (Nucl. Phys. A); Prog. Theor. Phys. 102, 1043 (1999). 5. S. Balberg and N. Barnea, Phys. Rev. C57, 409 (1998). 6. W.D. Langer and L. Rosen, Astrophys. and Space Sci. 6, 217 (1970). 7. V.R. Pandharipande, Nucl. Phys. A178, 123 (1971). 8. N.K. Glendenning Phys. Lett. B114, 392 (1982); Nucl. Phys. A 4 9 3 , 521 (1989). 9. F. Weber and M.K. Weigel, Nucl. Phys. A505, 779 (1989). 10. Y. Sugahara and H. Toki, Prog. Theor. Phys. 92, 803 (1994). 11. F. Schaffner and I.N. Mishustin, Phys. Rev. C53, 1416 (1996). 12. S. Balberg and A. Gal, Nucl. Phys. A265, 435 (1997). 13. M. Baldo, G.F. Burgio and H.J. Schultz, Phys. Rev. C58, 3688 (1998). 14. T. Takatsuka and R. Tamagaki, Prog. Theor. Phys. Suppl. N o . 1 1 2 , 27 (1993). 15. M.M. Nagels, T.A. Rijken and J.T. de Swart, Phys. Rev. D15, 2547 (1977). 16. T. Ueda, K. Tominaga, M. Yamaguchi, N. Kijima, D. Okamoto, K. Miyagawa and T. Yamada, Prog. Theor. Phys. 99, 891 (1998). 17. R.V. Reid, Ann. Phys. 50, 411 (1968). 18. T. Takatsuka, Prog. Theor. Phys. 50, 1754 (1973). 19. Y. Yamamoto, S. Nishizaki and T. Takatsuka, Prog. Theor. Phys. 103, No.5 (2000). 20. R.C. Malone, M.B. Johnson and H.A. Bethe, Astrophys. J. 199, 741 (1975). 21. S. Shinmura, I. Arisaka, K. Nakagawa and M. Wada, inpreparation. 22. R. Tamagaki, Prog. Theor. Phys. 39, 91 (1968).

Ferromagnetism of quark liquid and magnetars

Department

of Physics, E-mail:

Toshitaka Tatsumi Kyoto University, Kyoto 606-8502, [email protected]

Japan

Spontaneous magnetization of quark liquid is examined on the analogy with that in electron gas. It is pointed out that quark liquid has potential to be ferromagnetic at rather low densities, around nuclear saturation density. Some comments are given as for implications on magnetars.

1

Introduction

Recently a new type of neutron stars with extraordinary magnetic field, usually called magnetars, has attracted much attention in connection with pulsars associated with soft-gamma-ray repeaters (SGR) and anomalous X-ray pulsars (AXP). There have been known several magnetar candidates such as SGR 1806-20 and SGR 1900+14 1. Various analysis including the P - P curve have indicated an intense magnetic field of O(10 14-1,r> ) G, while ordinary radio pulsars have a magnetic field of O ( 1 0 1 2 _ u ) G . The origin of the strong magnetic field in compact stars, especially neutron stars, has been an open problem. Recent discoveiy of magnetars seems to renew this problem. Conservation of the magnetic flux during the collapse of a main sequence star has been a naive idea to understand the magnetic field in neutron stars 2 . Then the strength B should be proportional to R~2, where R is a radius of a star; for example, the sun, a typical main sequence star, has a magnetic field of O(10'j)G with the radius R ~ 10 1 0 - 1 1 cm. By decreasing the radius to 106cm for neutron stars B = O(10 12 )G, which is consistent with observations for radio pulsars. However, if this argument is extrapolated to explain the magnetic field for magnetars, we are lead to a contradiction: their radius should be O(104)cni to get an increase in B by a factor of ~ 10 12 , which is much less than the Schwartzschild radius of neutron stars with the canonical mass M = 1.4M 0 , RSc.h = IGM/c2 = 4 x 105cm. When we compare the energy scales for systems such as atomic system ( e - ) , nucleon system (p) and quark system (q), we can get a hint about the origin of the magnetic field. In Table 1. we list the interaction energy, Pint = ViB, of the magnetic field B = 1015G and each constituent with the Dirac magnetic moment, m = eiti/2miC. We also Ust a typical energy scale EtyP for each system. Then we can see that Etyp Eint for the nucleon and the quark systems; that is, the strength of O(10 l s )G is very large for the former system with the elec-

313

314

7(i;[MeV] Eint[UeV} Etyp

f Ol 5- 6 O(KeV)

p W 2.5 x l O " 3 > O(MeV)

q TTOO 2.5 x 10~ 2 - 2.5 > O(MeV)

Table 1.

tromagnetic: interaction, while it is not large for t h e latter systems with the strong interaction. Hence it m a y be conceivable t h a t the strong interaction should easily produce the magnetic field of the above magnitude. Since there is a bulk hadronic matter beyond nuclear saturation density (?io ~ 0.16fm~ 3 ) inside neutron stars, it should be interesting to consider the hadronic origin of the magnetic field; ferromagnetism or spin-polarization of hadronic m a t t e r may give such magnetic field. In 70's, j u s t after the first discovery of pulsars, there have been done many works about the possibility of the ferromagnetic transition in dense neutron matter, using G—matrix calculations or variations! calculations with the real­ istic nuclear forces. Through these works there seems to be a consensus t h a t ferromagnetic phase, if it exists, should be at very high densities, and there is no transition at rather low densities relevant to neutron stars i . We consider here the possibility of ferromagnetism of quark liquid interact­ ing with the one-gluon-exchange (OGE) i n t e r a c t i o n 4 . One believes t h a t there are deconfinement transition and chiral symmetry restoration at finite baryon density, while their critical densities have not been fixed yet. One interesting suggestion is t h a t three-flavor symmetric quark m a t t e r (strange m a t t e r ) may be the true ground state of QCD at finite baryon d e n s i t y 5 ' 6 . If this is the case, quark stars (strange quark stars), can exist in a different branch from the neutron-star branch in the mass-radius p l a n e 7 . Usually one implicitly assumes t h a t the ground state of quark m a t t e r is unpolarized. We examine here the possibility of polarization of quark m a t t e r . We shall see our results should give an origin of the strong magnetic field for magnetars in the context of strange quark-star scenario. 2 2.1

S p o n t a n e o u s m a g n e t i z a t i o n of q u a r k l i q u i d Rp.lativistic

formulation,

Quark liquid should be totally color singlet (neutral), which means t h a t only the exchange interaction between quarks is relevant there. This may remind us of electron system with the Coulomb force in a neutralizing positive charge

315

background. In 1929 Block first suggested a possibility of ferromagnetism of electron system 8 . He lias shown that there is a trade oif between the kinetic and the exchange energies as a function of a polarization parameter, the latter of which favors the spin alignment due to a quantum effect; electrons with the same spin orientation can effectively avoid the Coulomb repulsion due to the Pauli exclusion principle. When the energy gain clue to tke spin alignment dominate over the increase in the kinetic energy at some density, the unpolarized electron gas suddenly turns into the completely polarized state. In the following we discuss the possibility of ferromagnetism of quark liquid on the analogy with electron gas (Fig. 1).

k

Figure 1: Exchange interactions for electrons with the Coulomb force (left) and quarks with OGE interaction (right).

It is to be noted that there is one big difference between them; quarks should be treated in a relativistic way. The concept of the spin orientation is not well defined in relativistic theories, while each quark has two polarization degrees of freedom. Here we define the spin-up and -down states in tke rest frame of each quark. Then the projector onto states of definite polarization is given by P(«) = (1 + ^(f)/2 with the 4-pseudovector a, k(c ■ k)

a° =

c + mq(Ek+mqy

C-k

(1)

in a

for a quark moving with the momentum k = (Ek, k) 9 . Tke 4-pseudovector a is reduced into the axial vector £ (|£| = 1) in the rest frame, which is twice the mean spin vector in the rest frame. Hence a or £ can specify the polarized state. The exchange interaction between two quarks with momenta k and q (Fig. 1) is written as 2 „2 (f m„ / k C qc'-9mqEk

m„ [2m EVa "

k

q

m\a ■

1 b] (k - q)*

(2)

316

t t t t t Figure 2: Heisenberg ferromagnet in the coordinate space (left) and quark ferromegnet in the phase space (right).

where the 4-pseudovector b is given by the same form as in Eq. (1) for the momentum q. The exchange energy is then given by the integration of the interaction (2) over the two Fermi seas a for the spin-up and -down states; eventually, it consists of two contributions, ^x = e™ n _ / H p + 4!!p-

(3)

The first one arises from the interaction between quarks with the same po­ larization, while the second one with the opposite polarization. The non-flip contribution is the similar one as in electron gas, while the flip contribution is a genuine relativistic effect and absent in electron gas. We shall see that this relativistic effect leads to a novel mechanism of ferroiiiagnetism of quark liquid. 2.2

Symmetry consideration of ferromagnetic phase

Usual Heisenberg model describes the spin-spin interaction between adjacent spins localized at lattice points; that is, the Heisenberg ferromagnet is the spin alignment in coordinate space. On the other hand, the concept of spin alignment in quark liquid requires an extension to the phase space because of the coupling of spin with momentum. Since the spatial part of the quark wave functions take the plane wave, the spin orientation is obviously uniform in coordinate space, once £ is given. On the other hand, the spin does not necessarily take the same orientation in momentum space: generally £ should be momentum dependent (see Fig. 2). The most favorite configuration in momentum space may be determined by an energetic consideration, while it seems to be a difficult task. We consider here only a naive case, where the spin orientation is uniform even in the phase space. This is a direct analog of the nonrelativistic version. a

W e , here, don't consider any deformation of Fermi spheres for simplicity, while they may be deformed in a realistic case due to the momentum dependent interaction.

317

Anyway, the ferromagnetic phase is a spontaneously symmetry broken state with respect to the rotational symmetry in coordinate space: the order parameter is the mean value of £, {£), and symmetry is broken from G = 0(3) to H = 0(2) once (£) takes a special orientation. 3

Examples

We show some results about the total energy of quark liquid, etot — (kin + c?.x, by adding the kinetic term €*,,•„. Since gluons have not the flavor quantum numbers, we can consider one flavor quark matter without loss of generality. Then quark number density directly corresponds to baiyon number density, if we assume the three flavor symmetric quark matter as mentioned in §1. There are two QCD parameters in our theory: the quark mass mq and the quark-gluon coupling constant ac. These values are not well determined so far. In particular, the concept of quark mass involves subtle issues; it depends on the current or constituent quark picture and may be also related to the existence of chiral phase transition 10 . Here we allow some range for these parameters and take, for example, a set, mq = 300MeV for strange quark and ac = 2.2, given by the MIT bag model 11 . In Fig. 3 two results are presented as functions of the polarization parameter p defined by the difference of the number of the spin-up and -down quarks, nf. — n~ — pnq. The results clearly show the first order phase transition, while it is of second order in the Heisenberg model. The critical density is around nq ~ 0.16fm~'5 in this case, which corresponds to ?io f° r flavor symmetric quark matter. Note that there is a metastable ferromagnetic state (the local minimum) even above the critical density. Magnetic properties of quark liquid are characterized by three quantities, fie,x and »/; fie = etot{p = 1) — etot(jj — 0), which is a measure for ferromagnetism to appear in the ground state. For small p -C 1, etot - e,.ot(p = 0) = x " V + 0(p4).

(4)

X is proportional to the magnetic susceptibility. In our case it is less relevant since the phase transition is of first order. Finally, // = detot/dp | P =i, which is a measure for metastability to to exist. In Fig. 4 we present a phase diagram in the mq — ac plane for nq = 0.3fm - ! , which corresponds to about twice IIQ for flavor symmetric quark matter. The region above the solid line shows the ferromagnetic phase and that bounded by the dashed and dash-dotted lines indicates the existence of the matastable state. For heavy quarks, which may correspond to the current s quarks or the constituent quarks before chiral symmetry restoration, the ferromagnetic

318 n a =0.3fm J m.=300 MeV, ^.=2.2 242 240

n„=0.2 (fm'3

238

? 7rE or KN —> 7rA are energetically almost closed for such a deeply-bound state. Kaon absorption by two nucleons (KNN —> YN) gives little width since two nucleons have to participate to the reaction. Even though the width is twice wider the Is state should be seen well separated since the next excited state (lp) is

324 expected to appear 40 MeV higher. 4. (K~,N)

reaction to excite kaonic nuclei

The (K~,N) reaction where a nucleon (N) is either a proton or a neutron is shown schematically in figure 1. The nucleon is knocked out in the forward direction leaving a kaon scattered backward in the vertex where the K + N —>■ K + N takes place. This reaction can thus provide a virtual K~ or K° beam which excites kaonic nuclei. This feature is quite different from other strangeness transfer reactions like (K~,ir), (TT±, K+) and (7, K+) extensively used so far. They primarily produce hyperons and thus are sensitive to states mostly composed of a hyperon and a nucleus.

,A-1

Figure 1 Diagram for the formation of kaonic nuclei via the (K~,N) reaction. The kaon, the nu­ cleon, and the nucleus are denoted by the dashed, thin solid and multiple lines, re­ spectively. The kaonic nucleus is denoted by the multiple lines with the dashed line.

The momentum transfer, which characterizes the reaction, is shown in figure 2. It depends on the binding energy of a kaon. We are interested in states well bound in a nucleus (BE = 100 ~150 MeV). The momentum transfer for the states is fairly large (q = 0.3 ~ 0.4 GeV/c) and depends little on the incident kaon momentum for PK = 0.5 ~ 1.5 GeV/c, where intense kaon beams are available. Therefore one can choose the incident momentum for the convenience of an experiment. It is a little beyond the Fermi momentum and the reaction has characteristics similar to the (ir+, K+) reaction for hypernuclear production where so-called stretched states are preferentially excited [13]. Figure 2 The momentum transfer of the (K~,N) reaction at 0 degrees is shown for four reactions. Here binding energy of kaonic nucleus ^Mg is taken to be 150 MeV.

p(K,p)K- d(K\n)A(1405) "Si(K\p)"Al+K-"SidCpJ^g -

> £L 300

1000

PK (MeV/c)

1200

1400

1600

325 Recently deeply bound -K~ atoms were observed by the (d,3 He) reaction [14]. A small momentum transfer (~ 60MeV/c) was vital to excite the atomic states which were typi­ cally characterized by the size of the atomic orbits. If one wishes to excite kaonic atoms, a momentum transfer less than 100 MeV/c is desirable. It is achieved by kaon beams less than 0.4 GeV/c where available beam intensity is very small. The repulsive nature of the 7r-nucleus interaction allows no nuclear state although the strong attractive .RT-nucleus potential makes kaonic nuclei exist. The (K~,N) reaction can excite the deeply bound kaonic nuclei with large cross section in spite of the large momentum transfer of the reac­ tion. For the excitation of nuclear states the momentum transfer is typically characterized by the Fermi momentum. The (K~, N) reaction on deuteron is the simplest reaction by which one can study the KN component of excited hyperons. The d(K~,p) reaction excites K~n states which can only have 1 = 1. On the other hand d(K~,n) reaction excites a K~p state which can have either / = 1 or / = 0. Cross sections to the excited hyperons depend on their KN component. For instance, the well known A(1405 MeV) should be abundantly excited by the [K~,n) reaction if it is a KN bound state with I = 0 as usually believed. The d(K~,p) reaction, in particular, gives information on the K~n interaction below the threshold, which plays decisive role on the kaon condensation in the neutron stars. 5. Formalism We adopt here the distorted wave impulse approximation (DWIA) to evaluate the cross section. The DWIA calculation requires (a) distorted waves for entrance and exit channels, (b) two body transition amplitudes for the elementary (K~,N) process, and (c) a form factor given by initial nuclear and kaonic-nuclear wave functions. Relevant formulas for the calculation can be found elsewhere [13]. The differential cross section in the laboratory system for the formation of kaonic nu­ cleus is given by ,

da

/ ,

\K-N^NK-

Ida \

dn = {dn)Lfl.

_.

,„.

N

(1)

°"-

It is given by the two body laboratory cross section multiplied by the so-called effective nucleon number (Neff). We first use the plane wave approximation to evaluate N^. At 0 degrees, where only non-spin flip amplitude is relevant, N^Jf is given by NlJ} = (2 J + 1) (2jN + 1) (2EK + l)(£«

5

f ) F{q).

(2)

In this equation we assumed that a nucleon in a jn orbit is knocked out and a kaon enters in an £K orbit making transition from 0 + closed shell target to a spin J state. Here the form factor F(q) is given by the initial nucleon and final kaon wave functions as F(q) = (Jr2drRK(r)RN(r)jL(qr))

,

where L = J ± | is the transferred angular momentum.

(3)

326 For an oscillator potential of radius parameter b, the radial wave function is

Mr) = c collimator was changed by 1mm steps, and 20Ne7+ ions yield was measured at the entrance and the focal point of RMS when the optical elements of RMS was tuned for AM=1 heavier ions than 20Ne ions. The result is shown in Fig.2. It shows that the ideal beam suppression factor becomes several x 10"5 and if the vertical beam position is 3mm lower than the center, it becomes near 10"3. For the determination of the contribution to the beam suppression factor by the second origin, the beam suppression factors were measured with 20Ne beams whose

337

D-mag

Target Chamber Electric Deflector

'%?

J^Fb-mag rin

Q

m

°0/'M

r

D_ma

^

C foil

9 -

^

Detector I H -Linac Chamber Fig. 1 The ion configuration from the linac to a detector system. At the downstream of a target chamber, There is a recoil mass separator.

beam suppression

olOl

•

u

factor

transprot

I

efficiency

no"

■

to

c o 0)

-4-

i

i

i

i

i

i

•

|

i

I

]

•

I

i

-1 •: 1 0

01 3 W TJ -t

O i—*

CD

—*

Q.

m

■

E

10'

CO* 3 O

CO CD

XI

cl . . . . . . . . i i . . . . i . . . . I . . . . j , , , , I , r , , i , , , ,

1 0 :

4

. 3 . 2 - 1 0 1 2 3 Collimator position [mm]

10

-3

Fig.2 The measured beam suppression factor with '"Ne7* beam. It was measured by changing vertical position of the collimator. RMS was tuned for AM=1 heavier ions than ones of the beam.

338

charge state were from 4+ to 8+. As the result, the beams whose charge state were lower than 4+ collided with the electric electrode, but their suppression factor were lower than 2x10"*. From the above mentioned fact, it was turned out that the main contribution to the bad beam suppression factor was that the beam collided with the vertical chamber wall in the dipole magnet. So, two new X-Y stairing magnets and more beam monitors were constructed and the vertical position of the ion optical elements were adjusted again. As the result, the beam suppression factor of RMS itself become 1 x 10-4. For the search of the resonance levels of l9Na by ,8Ne(p,p) reaction, required beam intensity would be 103pps. Because the 18Ne beam would be directly transported to a detector system including a gas target, consisting of a multiwire proportional counter and a position sensitive silicon detector. The l8Ne beam intensity had become 105 pps at the exit of ISOL, and contaminant ls O stable nuclear beam had become a few nA at the exit of ISOL. So, we had planned to separate ^Ne9* and 18 0 whose maximum charge state was 8+. The ions with A/q=18/2+ would be accelerated up to 1.04MeV/u. After passing through a thin carbon foil(10u.g/cm2) at the exit of the linac, the ions with A/q=18/9+ would be analyzed by available D-D-E-D. The measured beam suppression factor to , 8 0 became 1 X 10"8. As the final result, ,8 Ne beam intensity at the detector became 3 x 103pps and the maximum yield ratio of contaminant ,s O ions to 18Ne became 1.5%. In summary, a recoil mass separator was constructed for nuclear astrophysics and tested its performance. The measured beam suppression factor of RMS itself became 1 x 10"4 when it tuned for AM=1 heavier ions than the beam ions with A~20. It was the upper limit to carry out the 19Ne(p,y) experiment. By a charge state breeding, the suppression factor to contaminant stable nuclear beams became 1 x 10'8. This method is available to purify radioactive nuclear beams.

References 1. 2. 3. 4. 5. 6.

T.Nomura, Nucl. Instr. Meth. 870(1992)407. M.Oyaizu, et al., Rev. Sci. Inst. 69-2(1998)770. M.Wada, et al., Nucl. Inst. Meth. Res. A337(1993J11 M.Tomizawa, et al., Heavy ion Accelarator Technology: 8th International Conference, edited by Kenneth. W.Shepard (AIP, New York, 1999)451. A.E.Champagne, M.Wiescher., Annu. Rev. Nucl. Part. Sci. 42(1992)39. R.D.Pages, et. al., Phys. Rev. Lett. 73(1994)3066.

Collision between Neutron Star and Axion Star as a Source of Gamma Ray Burst and Ultra High Energy Cosmic Ray A. Iwazaki of Physics, Nishogakusha University, Chiba 277-8585, Japan E-mail: [email protected]

Department

We propose a model in which the ultra high energy cosmic rays and gamma ray bursts are produced by collisions between neutron stars and axion stars. An ac­ celeration of the cosmic ray is done by an electric field, ~ 10 15 eV c m " 1 , which is induced in the axion star by the strong magnetic field > 10 12 G of the neutron stars. On the other hand, gamma ray bursts are generated in the collisions between axion stars and neutron stars with relatively small magnetic field, e.g. ~ 10 10 G. In the former case the axion star evapolates emitting ultra high energy cosmic rays before colliding directly with the neutron stars, while in the latter case the axion star collides directly with the neutron star and dissipates rapidly its whole energy in an outercrust of the neutron star, which leads to a gamma ray burst. To explain both phenomena we need to assume the mass of the axion such as 1 0 - 9 eV. With this choice we can explain huge energies 10 s 4 erg of the gamma ray bursts as well as the ultra high energies ~ 10 2 0 eV of the cosmic rays. Additionally, it turns out that these axion stars are plausible candidates for MACHOs.

1

Introduction

The ultra high energy cosmic rays ( UHECRs ) is one of most mysterious phenomena in astrophysics*. We do not still have a reliable generation mech­ anism of the cosmic rays with extremely high energies > 1020 eV. Similarly, the gamma ray burst is one of most mysterious phenomena in astrophysics although some plausible generation mechanisms are proposed. The difficulty for explaining both phenomena is how huge energies are released; energies of the cosmic rays, > 1020 eV, and energies of gamma ray bursts, ~ 1052 erg. On the other hand, the dark matter in the Universe 2 is one of most mysterious puzzles in cosmology. Axion 3 ' 4 is one of the most plausible candidates for the dark matter although even its existence has not yet been confirmed. 'Proba­ bly, some of axions with mass ma may form boson stars ( axion stars ) in the present Universe by gravitational cooling 5 or gravitational collapse of axion clumps formed at the period of QCD phase transition 6 . Here we wish to sketch our results that collisions between axion stars and neutron stars generate UHECRs and GRBs. In our model the collision between the axion star and the neutron star with strong magnetic field > 1012 G produces both UHECRs and GRBs with very short durations ( less than

339

340

millisecond ) and very hard gamma rays. The collision between the axion star and the neutron star with relatively weak magnetic field ~ 1010 G produces only GRBs with both of short and long durations in this case; UHECRs can not be produced. In order to derive the results we need to assume that the mass of the axion is given by ~ 10~ 9 eV. Then, we have an additional bonus that the axion stars are plausible candidates for MACHOs 7 since their masses are given by ~ 1O _ 1 M 0 ; M© is solar mass. The essence in our mechanism 8 is that the axion star induces an electric field when it is exposed to the magnetic field of the neutron star. The strength of the field Ea is proportional to the strength of the magnetic field B, Ea ~ 1015 eV c m ' 1 B12mg, where B12 = B/1012 G and m 9 = m 0 / l ( r 9 eV with ma denoting the mass of the axion. This electric field can accelerate charged particles so that they can gain the energies 1020 eV when the strength of the magnetic field is given by ~ 1012 G. The strong electric field, however, is unstable 9 against electron-positron pair creations and hence the axion star decays producing UHECRs before colliding directly with the neutron star. It means that the axion star evaporates very rapidly around the neutron star with its surface magnetic field B > 1012 G. On the other hand, the electric field dissipates its energy in the conducting medium of the neutron star when its magnetic field is given by ~ 1010 G. It turns out that in such a case the electric field is stable so that the axion star can collide directly with the neutron star. Since its whole energy ~ 10 53 erg is dissipated only in an outercrust of the neutron star, the ejection leading to GRB is composed of particles in the crust. Thus, the baryonic contamination of the ejection is less than 1O~5M0 which is required observationally. 2

Axion Star

The axion star is a coherent object of the real scalar field a(x) describing the axion. An approximate form of the solution 10 representing axion star is given such that a(x) = fpQdo sin(m a i) exp(—r/Ra) where t ( r ) is time ( radial ) coordinate and JPQ is the decay constant of the axion. The value of JPQ is constrained 2 conventinally from cosmological and astrophysical considera­ tions 2 ' 4 such that 1010 GeV < fpQ < 1012 GeV (the axion mass ma is given in terms of JPQ such that ma ~ 107 GeV2/fpQ ). However, when we assume un­ conventionally entropy productons below the temperature 1 GeV of the early Universe, we may be released from the constraints 11 . In this paper we assume that fPQ ~ 1016 GeV or ma ~ 10" 9 eV. In the formula, Ra represents the radius of the axion star numerically given in terms of mass Ma of the axion star, Ra ~ 1.6 x i0 5 cmm^~ 2 10~1MQ/Ma.

341

Similarly, the amplitude ao in the solution is represented such that oo = 1.73 x 10 2 (10 5 cm/i? o ) 2 l ( r 9 e V / m a . Therefore, we find that the solution is parameterized by one free parameter, either one of the mass Ma or the radius Ra of the axion star. It is also important to note that the solution is not static but oscillating with the frequency of ma/2ir. The mass of the axion star may be typically given by the critical mass M c of the axionic boson star, Mc ~ 10 _ 1 M Q 10 _ 9 eV/m a ; axion stars with masses larger than the critical mass collapse gravitationally into black holes. Thus, the corresponding radius Ra for this critical mass is given such that Ra ~ 1.6 x H ^ m ^ c m . Let us explain how an electric field is induced in the axion star when it is under the magnetic field of a neutron star. Owing to the interaction between the axion and the electromagnetic field described by Lail = caaE • B/fpQ7r, where the value of c is of the order of unity, the Gauss law is modified such that dE = — cad ■ (aB)/fpQir + "matter" where the last term "matter" denotes contributions from ordinary matters. The first term in the right hand side represents the contribution from the axion. Thus, substituting the above solution into a(x) in the Gauss law, we find that the electric field induced in the axion star, Ea = —caa(x)B/fpQTr ~ 1015 eV c m - 1 Bumg with a ~ 1/137. This electric field is oscillating with the frequency, ma/2ir ~ 2.4x 105 ma Hz. Thus a charged particle with charge Ze can be accelerated in a direction within a half of the period, n/ma or in a distance ~ Ra ( ~ K/ma x light velocity ) by this field. Thus, the energy AE obtained by the particle is given such that AE = ZeEa x Tr/ma x light velocity ~ 10 2 0 ZeVSi 2 Therefore, the electric field of the axion stars can accelerate the charged particle so that its energy reaches ~ 10 20 Z eV of UHECRs. As we have mentioned, this strong electric field is unstable so that it decays very rapidly into electron-positron pairs, which are converted to baryons and photons after their production. These are UHECRs in our model. On the other hand, the electric field is stable when the magnetic field of the neutron star is not so strong, e.g. ~ 1010 G. Thus, the axion star collides directly with the neutron star and dissipate its whole energy ~ 10 53 erg in the conducting medium of the outercrust. The rate of the dissipation has been estimeted such as 10 46 erg/scm 3 , while the energy density of the axion star is given by 10 38 erg/cm 3 . Thus it turns out that the dissipation is very rapid; the axion star never enter a core of the neutron star. This rapid dissipation generates jets formed by particles of the neutron star, which leads to GRBs; the jets with small solid angles are produced since the ejections are accelerated by the strong electric field ~ 10 13 B i 0 e V c m _ 1 . In these ways UHECRs and GRBs are generated by the collisions between the axion stars and the neutron

342 stars. Acknowledgments This work is supported by the Grant-in-Aid for Scientific Research from the Ministry of Education, Science and Culture of Japan No. 10640284 References 1. M.A. Lawrence et al. J. Phys. G. Nucl. Part. Phys. 17, 773 (1991), D.J. Bird et al. Phys. Rev. Lett. 71, 3401 (1993); ApJ 424, 491 (1994), N. Hayashida et al. Phys. Rev. Lett. 73, 3491 (1994), M. Takeda et al. Phys. Rev. Lett. 81, 1163 (1998); ApJ, 522, 225 (1999). 2. For a review, see, for example, E.W. Kolb and M.S. Turner, The Early Universe, Addison-Wesley, New York, (1990). 3. R.D. Peccei and H.R. Quinn, Phys. Rev. Lett. 38, 1440 (1977), S. Weinberg, Phys. Rev. Lett. 40, 223 (1978), F. Wilczeck, Phys. Rev. Lett. 40, 279 (1978). 4. J.E. Kim, Phys. Rep. 150, 1 (1987). 5. E. Seidel and W.M. Suen, Phys. Rev. Lett. 72, 2516 (1994). 6. E.W. Kolb and I.I. Tkachev, Phys. Rev. Lett. 71, 3051 (1993); Phys. Rev. D49, 5040 (1994). 7. C. Alcock et al. Astrophys. J. 449, 28 (1995), for a review, see C.S. Kochanek and J.N. Hewitt, eds, "Astrophysical Applications of Gravitational Lensing", IAU Symp. 173 (1996) (Kluwer, Dordrecht). 8. A. Iwazaki, Phys. Lett. B455, 192 (1999); Phys. Rev. D60 025001 (1999); Prog. Theor. Phys. 101, 1253 (1999). 9. J. Schwinger, Phys. Rev. 82, 664 (1951). 10. E. Seidel and W.M. Suen, Phys. Rev. Lett. 66, 1659 (1991), A, Iwazaki, Phys. Lett. B451, 123 (1999). 11. P.J. Steinhardt and M.S. Turner, Phys. Lett. B129 51, (1983).

Double B e t a Decays of

100

M o b y E L E G A N T V at O t o C o s m o Observatory

N. Kudomi a , H. Ejiri a , K. Fushimi b , K. Hayashi a , T. Kishimoto c , K. Kume a , H. Kuramoto a , H. Ohsumi c , K. Takahisa a , Y. Tsujimoto a , and S. Yoshidaa a)RCNP, Osaka Urtiv., Ibaraki, Osaka 567, Japan b)Facul. of Integ. Arts and Sci., The Univ. of Tokushima, Tokushima, 770, Japan c)Dept. of Phys., Osaka Univ., Toyonaka, Osaka 560, Japan Exclusive measurements of neutrino-less double beta decays(0^/3/3) of 1 0 0 M o were made by means of ELEGANT V. The present status of the double beta decay experiment with ELEGANT V is presented. The data at Oto lab., being combined with the data at Kamioka, gives stringent limits on half-lives for 0vf3j3 and (mv) < 1.8eV for the Majorana neutrino mass.

Double b e t a decays are of current interest from b o t h astroparticle and nuclear physics view p o i n t s 1 . T h e Ov/3/3, which violate t h e lepton number con­ servation law, provide one with very sensitive tests for t h e Majorana neutrino mass, t h e right handed weak currents and so on. 2i//3/3 gives directly the nuclear matrix element?. This is used t o verify nuclear structure calculations and t o se­ lect appropriate spin-isospin interactions HTa t o be used for evaluating t h e nu­ clear matrix element for Qvf3f3. Since /3/3 transition rates depend largely on t h e nuclear matrix elements, it is i m p o r t a n t t o study /?/? on several nuclei in order to extract universal values of physics quantities. T h e transition rate for lv/3/3 is simply written in terms of t h e M2v as [ T ^ ] - 1 = G2v\M2v\2, where T2J2 is t h e half-life and G2v is t h e phase space factor. T h u s M2v is derived experimentally from t h e observed 21/(3/3. T h e transition rate for Ov/3/3 associated with t h e neu­ trino exchange is written in t e r m s of the neutrino mass term (m„) and the R H C terms of (A) and (r,) a s 1 , [^(v)]'1 = G*v\M°v\2((mv) + C A (A) + Cn{r))f, 0v 0v where G and M are the phase space factor and t h e nuclear matrix element for t h e mass term. C,, with i being m, A and r), are t h e nuclear responses in units of M0v. These 01/(3/3 transition rates include both t h e nuclear matrix el­ ements (form factors) relevant to nuclear structures and t h e physics quantities relevant to particle physics. So far, Ov/3/3 decay rates have been studied on several nuclei, unique fea­ tures of this work of 1 0 0 Mo with E L E G A N T V4 are as follows. (1) Energy and angular correlations of two (3 and j are measured by E L E G A N T V. T h u s limits on the Ov/3/3 for individual terms can be obtained. (2) 1 0 0 M o has large phase space factors and the 2v/3/3 m a t r i x element is k n o w n 3 . (3) Origins of

343

344

/

00

\x —A< /

—^

|

Mo Source film Lead Shield Copper Shield

NQ

-*iB

Upper-Nal

3

a"T I

^

Lower-Nal

.. i n m n n

«

Chamber - B

■« *

Chamber - A Lower-PL

innnn

Figure 1: Schematic view of ELEGANT V 4 .

the background are well investigated by the f3 and 7 correlation. Thus the correction for them is possible. The obtained limits on (mv) and others, depending somewhat on the ma­ trix elements used, are most stringent for 100 Mo. They are same orders of magnitudes as the values derived from other nuclei. Thus the present data with the unique points as given above, together with other data, may give stringent limits on the relevant values beyond the standard theory. ELEGANT V consists of drift chambers for /? trajectories, plastic scintillators(PL) for f3 ray energies and arrival times, and Nal scintillators(Nal) for 7 and X rays4(Fig. 1). The total weight of 100 Mo is 171gr. PL's and Nal's are calibrated by 7-rays of 511keV and 1275keV from the 22 Na 5 . The detection efficiencies of D C s are checked by the /?-ray from 90 Sr passing through DCs. At present, the j3/3 decays of 100 Mo are measured at Oto Cosmo Obser­ vatory with 1200m.w.e.. The sum energy spectrum is obtained by selecting events with several conditions with live time for the 6766hrs(Fig. 2). The major part of the spectrum is the 2z//?/? component. Background contribu­ tions from the natural radioactive contaminations were estimated as follows. The 214 Bi(Q^ = 3.28MeV) and 208 Tl(Q /3 = 4.99MeV) are two major isotopes, which may give background events in the 0^/?/? energy window. These isotopes are decay products of 238 U- and 232 Th-chain isotopes. 214 Bi is produced also from Rn contained in the air. As the first step, the total amounts of them were evaluated from the single electron//3-ray event rates in coincidence with 7-rays characteristic of the decays of each contents(Table 1). On the basis of the estimated 214 Bi and 208 T1 contents, Monte Carlo cal­ culations were carried out to evaluate the fake event rates by these isotopes, which may survive after the selections. The observed spectrum is reproduced

345

2

2.2

2.4

2.6

2.8

3 3.2 3.4 0 sum •n*rgy(M«V)

Figure 2: The sum energy spectrum.

by the sum of the estimated BG spectrum and the 2vj3j3 spectrum with the previously known half-life of 1.15 x 10 19 y 3 . There are no excess Of/?/? counts beyond the statistical fluctuations of the 2i>/?/?-|-BG events. In order to obtain upper limits on the number of the counts, the detection efficiencies for individ­ ual processes of the (m,), (A) and (77) terms were evaluated by Monte Carlo simulations, as shown in Table 2. The limits on the half-lives are obtained from the number of the observed counts and the number of the estimated 2z//?/?+BG counts. The standard method of the likelihood analysis was used 8 . The results at Oto together with those of the previous work at Kamioka are summarized in Table 2. Using the nuclear matrix elements 7 , the half-life limits leads to the upper limits of physics quantities. The present work measures the energy and angular correlations of two /? rays, which are important for the study of individual double beta processes. It is also important to study the double beta decay of several nuclei, because

Table 1: The

Origin * 14 Bi

Location Source DC-gas PL

214

Bi and

208

T 1 contents.

Amount (8.3 ± 1.7) x 10- 3 (2.2 ± 0.5) x 10-* (6.5 ±1.5) x 10-*

Bq/kg Bq/nr 5 Bq/m

346 Table 2: Limits with 68(90)%CL on the half-lives for the Oi//30 decay of and those obtained by combining with the Kamioka limits.

100

M o at Oto lab.,

(Oto(Preliminary), Live Time=:6766hrs) E window (MeV) 2.7-3.3 2.7-3.3 2.7-3.3 2.5-3.0

mode (mv) term (A) term {rj} term emission

0+ - * 0 +

Majoron

Yield (counts) 2 2 2 14

2v/3(5+BG (counts) 2.1 2.1 2.1 14.

e 0.18 0.11 0.16 0.029

Tl/2 (10 21 y) >46(25) >29(16) >40(22) >3.6(2.3)

(Combine) mode

0+^0+ Majoron

{mv) term (A) t e r m {rj) t e r m emission

T\/2 (10 2 1 y) 82(45) 50(27) 76(41) 6.9(4.0)

physics quantity (m„) 4M : j '4 : ' 1 ;'■ .'i

:

4000

M / MQ 6000

VAT/

8000

-.,_

V 0<m/sl

.-56M

*He V

•H

Figure 1: The isotopic composition of the ejecta of the strong aspherical explosion (£?K = 1.0 x 10 5 2 ergs, initial velocity ratio 8:1) in the direction of the jet (z:top) and perpendicular to the jet (r:bottom). The ordinate {Mr) indicates the initial spherical Lagrangian coordinate ( M r ) of the test particles. The expansion velocities of the particles are also shown.

351

b.'

ISdeg 30 deg Spherical

1}

J

J

t*

V

4700 4800 4900 5000 5100 5200 5300 5400 5500 5600 5700

5800 5900 6000 6100 6200 6300 6400 6500 6600 6700

Figure 2: The line proHles of the Fe blend (left) and of O I] 6300, 6363A (right) viewed from 15° (solid) and 30° (dashed) to the jet direction in the strong aspherical explosion model. For comparison, the spherical model (dotted) is also plotted.

5

Black Hole Binary GRO J1655-40

Recently, Keck observations reported interesting abundance features in the atmosphere of the secondary star of the black hole binary GRO J1655-40 4 , which consists of a massive black hole and a low mass companion. In the secondary, the abundance ratios of O, Mg, Si, S and Ti relative to H are almost ten times the solar ratios, while the Fe/H ratio is almost solar. It has been suggested that these a-elements were ejected by the supernova explosion of the primary and captured by the secondary 4 . In order to eject a large amount of Ti, S and Si and yet to leave a massive black hole remnant (which implies a large mass cut), explosive nucleosynthesis should take place in outer layers with sufficiently large Mr and low densities, so that the explosion should be as energetic as hypernovae 4 ' 12 . The small Fe enhancement can be explained by appropriately placing the mass cut in the spherical model 12 . The enhancement of Ti (without ejecting much Fe) is difficult to explain (see below), but to explain the large [Ti/Fe] in metal-poor stars is a well-known problem for core collapse supernova models. An alternative explanation is that the explosion of the primary was as­ pherical and little Fe was ejected in the direction of the secondary. Then it is likely that the secondary star captured material ejected along the r-direction (i.e. on the orbital plane), which contains relatively little Fe compared with the z-direction (see Fig. 1). We compare our nucleosynthesis calculations with

352

the observed abundance pattern. Our main results are as follows. (1) The observed enhancement of O, Mg, Si and S can be explained with the r-direction ejecta abundances in the hypernova (1052 ergs) explosion of a 16 M Q He star. (2) A large amount of Ti and a small amount of Fe are difficult to reproduce concurrently by our models. Ti (Cr) is produced deep in the supernova, where Fe (Ni) is also produced. To get much Ti and little Fe, the mass cut should be located at the interface between complete and incomplete Si burning (see Fig. 1). Even in that case, though, the amount of Ti predicted by the model is smaller than the observed value by a factor of 2. (3) Aspherical models have one advantage compared with a spherical model. In a spherical model, to reproduce the small amount of Fe observed in the secondary, the total amount of ejected 56 Ni should be very small (about 1 0 - 2 M 0 or less). This is much smaller than in other observed hypernovae 5,6 . Also, this does not seem to follow the observational tendency that the ejected 56 Ni mass is larger for more massive progenitors u . In contrast, if the explo­ sion was aspherical, a lot (more than 0.1 MQ) of Fe was ejected along the jet direction, so that total amount of Fe is not so different from other hypernovae. Although the Ti/Fe problem is yet to be solved, we favour the scenario that the progenitor of GRO J1655-40 was a very massive C+O (or He) star which underwent an aspherical hyper-energetic explosion and produced a large amount of Fe. 6

Conclusions

We have calculated the nucleosynthesis in aspherical hypernova explosions to compare with observations. We have succeeded to explain the unusual widths of the 0 and Fe nebular lines in SN1998bw with the strongly aspherical explo­ sion model. We have also shown that an aspherical hypernova explosion can explain the abundance features of the secondary star of the black hole binary GRO J1655-40 better than the spherical explosion model. This is the first step quantitatively to examine the observed character­ istics of hypernovae/GRB candidates using aspherical explosion models. The inferred asphericity gives us clues to the GRB central engines and the explosion mechanism of very massive stars 8 ' 7 . This work has been supported in part by the grant-in-Aid for Scientific Research (07CE2002,12640233) of the Ministry of Education, Science, Culture, and Sports in Japan.

353

References 1. Danziger, I.J., et al. 1999, in The Largest Explosions Since the Big Bang: Supernovae and Gamma Ray Bursts, eds. M. Livio., et al. (Baltimore: STScI), 9 2. Galama, T.J. et al. 1998, Nature, 395, 670 3. Hachisu, I. et al. 1991, ApJ, 368, L27 4. Israelian, G. et al. 1999, Nature, 401, 142 5. Iwamoto, K. et al. 1998, Nature, 395, 672 6. Iwamoto, K., et al. 2000, ApJ 534, 660 7. Khokhlov, A.M., et al. 1999, ApJ, 524, L107 8. MacFadyen, A.I. & Woosley, S.E. 1999, ApJ 524, 262 9. Nagataki, S. et al. 1997, ApJ, 486, 1026 10. Nomoto, K. & Hashimoto, M. 1988, Phys. Rep., 256, 173 11. Nomoto, K. et al. 2000, in "Supernovae and Gamma Ray Bursts" eds. M. Livio., et al. (Cambridge University Press) 12. Podsiadlowski, Ph., Nomoto, K., Mazzali, P.A., & Schmidt, B.P. 2000, preprint 13. Thielemann, F.-K., Nomoto, K., & Hashimoto, M. 1996, ApJ, 460, 408 14. Woosley, S.E., Eastman, R.G., k Schmidt, B.P. 1999, ApJ, 516, 788

Effects o f J e t - l i k e E x p l o s i o n i n S N 1 9 8 7 A

S. N A G A T A K I Department

of Physics, School of Science, the University 7-3-1 Hongo, Bunkyoku, Tokyo 113, Japan E-mail: [email protected]

of

Tokyo,

We studied the effects of jet-like explosion in SN 1987A. Calculations of the explo­ sive nucleosynthesis and the matter mixing in a jet-like explosion are performed and their results were compared with the observations of SN 1987A. It was shown that the jet-like explosion model is favored because the radioactive nuclei 4 4 T i is produced in a sufficient amount to explain the observed luminosity at 3600 days after the explosion. This is because the active alpha-rich freezeout takes place be­ hind the strong shock wave in the polar region. It was also shown that the observed line profiles of Fe[II] are well reproduced by the jet-like explosion model. In partic­ ular, the fast moving component travelling at (3000-4000) k m / s is well reproduced, which has not been reproduced by the spherical explosion models. Moreover, we concluded that the favored degree of a jet-like explosion to explain the tail of the light curve is consistent with the one favored in the calculation of the matter mix­ ing. The concluded ratio of the velocity along to the polar axis relative to that in the equatorial plane at the Si/Fe interface is ~ 2 : 1. This conclusion will give good constraints on the calculations of the dynamics of the collapse-driven supernova. We also found that the required amplitude for the initial velocity fluctuations as a seed of the matter mixing is ~ 30%. This result supports that the origin of the fluctuations is the dynamics of the core collapse rather than the convection in the progenitor. The asymmetry of the observed line profiles of Fe[II] can be explained when the assumption of the equatorial symmetry of the system is removed, which can be caused by the asymmetry of the jet-like explosion with respect to the equa­ torial plane. In the case of SN 1987A, the jet on the north pole has to be stronger than that on the south pole in order to reproduce the observed asymmetric line profiles. Such an asymmetry may also be the origin of the pulsar kick. When we believe some theories that cause such an asymmetric explosion, the proto-neutron star born in SN 1987A will be moving in the southern part of the remnant.

1

Introduction

Supernova is the site where heavy nuclei are synthesized. There are two types of supernova explosion. One is the thermonuclear explosion and the other is the collapse-driven explosion. In this study, we discuss the collapse-driven supernova explosion. There is a problem with the collapse-driven supernova. In spite of many excellent studies, its mechanism has not been understood completely. In fact, almost all of the numerical simulations failed in exploding the progenitor. Al­ though Wilson (1985) proposed the promising mechanism of 'delayed explosion' in his one-dimensional calculation, its role seems controversial 2 .

354

355

To solve the problem, some researchers investigated the effect of rota­ tion of the progenitor 3 . They performed multi-dimensional hydrodynamical simulations and reported the dynamics becomes quite different from spherical explosion when the effect of rotation is included. For example, Yamada & Sato (1991) reported the possibility that a jet-like explosion could occur. So, rotation may be a key process of a collapsing star and jet-like explosion may be a common phenomenon among collapse-driven supernovae. If this is true, nucleosynthesis in a collapse-driven supernova has to be recalculated since it has been calculated assuming a spherical explosion. In this study, nucleosynthesis and matter mixing in an axisymmetrically deformed supernova are investigated and their results are compared with the observations of SN 1987A. As a result, we show in this study that the degree ~2:1 is favored for the axisymmetrically deformed shock wave in SN 1987A. 2 2.1

Explosive Nucleosynthesis Models

Since there is still uncertainty as to the mechanism of a collapse-driven super­ nova, precise explosive nucleosynthesis calculations have not been performed from the beginning of core collapse. Instead, explosion energy is deposited artificially at the innermost boundary 5 . In this study, this method is taken and the explosion energy of 1.0 x 1051erg is injected at the Fe/Si interface. The initial velocity of matter behind the shock wave is assumed to be radial and proportional to r x +a^?^—', where r, 9, and a are radius, the zenith angle, and the free parameter which determine the degree of the axisymmetric explosion. In the present study, we take a — 0 for a spherical explosion and a = | , | , and | (these values mean that the ratios of the velocity are 2:1, 4:1, and 8:1, respectively) for the axisymmetric ones. We name these models SI, Al, A2, and A3. We assumed that the distribution of thermal energy is same as the velocity distribution and that total thermal energy is equal to total kinetic energy. 2.2

Results

The required amount of 44 Ti in order to reproduce the tail of the light curve in SN 1987A 78 9 and calculated ones are shown in Figure 1. The required amount of 44 Ti from the observation depends on its half-life, which is not determined completely. Because of this reason, the required amount is drawn as a function of its half-life. However, in spite of the difficulty in determining the half-life of the long-lived nuclei, recent experiments give a converged value, r ~ 60

356 1 I

o

S3

1

1

1

1

,

1

t[

1

1 1

A2

■

/ < >

J ^

Al , a2 u C a T ° •o 3

1

Degree of Jet-like Explosion

r a

1

1

'

i

.-—

SI • Wi XI

-

\/

0)

u

'3

o* a> o

i

40

,

,

.

.

,

,

i

,

,

,

i

,

60 80 100 Half-life of "Ti (yrs)

Figure 1: Calculated masses of 44 Ti by the SI, Al, and A2 models. The synthesized mass of 44 Ti becomes larger along with the degree of the jet-like explosion. Required amounts of 44 Ti are also shown as a function of its half-life (Mochizuki & Kumagai 1998; Mochizuki et al. 1999a; Kozma 1999; Lundqvist et al. 1999). The most reliable value for its half-life is ~ 60 yrs (Ahmad et al. 1998).

yrs. Calculated masses of 4 4 Ti in the models Si, Al, and A2 are represented as horizontal lines. We can see the tendency that the synthesized mass of 44 Ti becomes larger along with the degree of axisymmetrically deformed shock wave. It is noted that the model Al is a good one to explain the amount of 44 Ti in SN 1987A. 3

Matter Mixing

Calculated velocity distributions of 56 Ni are shown in Figure 2. Velocity dis­ tributions are calculated assuming that the angle between the line of sight and the symmetry axis is 44°, which is inferred from the form of the ring around SN 1987A 10 . As can be seen from Figure 2, fast moving component can not be reproduced in the spherical explosion models, which is consistent with other works n 12 . On the other hand, fast moving component is reproduced in the jet-like explosion models when the amplitude of the initial fluctuations is set to be 30%. We conclude that the velocity distribution of the model Al with the initial fluctuation of 30% is most similar to the observed one among the all models.

357 I ' ' ' I

I

T1

Model SI

0%

M.C. = S

5%

I

I

i \ I

1

1

T^

1

1

1 1

Model Al M.C. = S

i

i

i

j

i

r

i

l

i

.

i

l

---•0% 5% 30%

30% x E

'"ii . ;"

j?i

1 : P.^ S'.rf

|i*i i 1

1 , r 0) stars • right panel : Young (tform > 4 Gyr) stars Here, Z is the metal fraction in stars. Metal-poor stars show the extended distribution since these stars formed at an early epoch and at high latitude. The old and metal-rich stars are concentrated in the galactic center. The young stars are located near the galactic plane, because the most recent star formation occurs in the disk. From these results, we conclude that we obtain the stellar system that is very similar to the Galaxy. The early evolution of our model have revealed that the star located near the center (bulge stars) formed during the sub-galactic merger. Because of the strong star burst induced by the mergers, the metallicity distribution function of the bulge stars becomes as wide as observed. Also the observed bi-modal metallicity distribution function is naturally explained with our chemical and dynamical model (Figure 3). From these results, we suggest that the Galactic bulge was formed through the sub-galactic clump merger in the proto-galaxy. References 1. 2. 3. 4. 5. 6. 7. 8.

Mihalas, D., k Binney, J., Galactic Astronomy (second edition) Eggen, O.J.; Lynden-Bell, D., k Sandage, A.R., 1962, ApJ, 136, 748 Searle, L., k Zinn, R., 1978, ApJ, 225, 357 Katz, N., 1992, ApJ, 391, 502 Nakasato, N., Mori, M., k Nomoto, K., 1997, ApJ, 484, 608 Steinmetz, M., k Miiller, E., 1994, AkA, 281, 97 Steinmetz, M., k Miiller, E., 1995, MNRAS, 276, 549 Sugimoto, D., Chikada, Y., Makino, J., et al., 1990, Nature, 345, 33

361

Figure 1. The total SFR as a function of time.

Figure 2. The projected particle positions at t = 5 Gyr for three components are shown. From the left panel, the metal-poor stars, the metal-rich and old stars, and the young stars (see text). The size of the panel is 20Kpc X 20Kpc.

3 a o

—f

»

>

*

i

* — > — ■ — > —

10 5 2 erg), the progenitors of which have 30-40 MQ on their main-sequence. 4) We compare our yields with those of the observed metaldeficient stars. The abundance patterns are not in good accord and it implies that pair-instability supernovae were not dominant in the early Galaxy. 5) We also calculate theoretical light curves of a pair-instability supernova and find that they are very luminous (L ~ 10 4 3 erg s - 1 ) because of the large production of 5 6 Ni.

1

Introduction

Pair-instability supernovae originate from the stars which are more massive than ~ 100M Q (e.g., Fowler & Hoyle 1964). At the center of such massive stars, the temperature is so high when the oxygen-rich core forms that there exist many photons which have the energy greater than the rest-mass energy of two electrons. Besides, the effective adiabatic index T is close to 4/3 be­ cause of a large fraction of radiation pressure. Thus, copious electron-positron pairs are created and F drops below 4/3; then the star becomes dynamically unstable and begins to collapse. The central temperature gets higher and the contraction leads to the onset of explosive oxygen burning. Finally, oxygen burns explosively releasing a large amount of nuclear energy. After V becomes larger than 4/3 again, the infall is stopped and the cores start to re-expand. It is believed that a larger number of such massive stars existed, in the early Galaxy, than in the present epoch (e.g., Tohline 1980, Hutchins 1976). Both the initial absence of any metal and the influence of the background radiation would tend to make the stars considerably more massive. Therefore,

362

363

very massive stars are likely to have affected the Galactic chemical evolution through pair-instability supernovae. Most of the previous studies for pair-instability supernovae were limited in a simple analysis with a small reaction network; therefore a detailed nucle­ osynthesis analysis has never been fully carried out. Thus, we are motivated to investigate the abundances of pair-instability supernova ejecta and their potential impact on the early Galactic chemical evolution. 2

Explosions & Nucleosynthesis

Presupernova models of 150 — 3OOM0 with no initial metal are used for our calculations (Umeda & Nomoto 2000; see also Umeda et al. 1999). They evolve through the hydrogen, helium and carbon burning stages, to the onset of the e + e"-pair creation instability. No mass loss is assumed because of no metals, near the surface. We use an equation of state which includes the effect of electron-positron pair creation to track the evolution correctly (Nomoto & Hashimoto 1988). Screening effects and Coulomb interaction (Slattery et al. 1982) are also taken into account. The effect of convection is neglected on the assumption that the evolution is too fast to allow for effective mixing (see Fraley 1968). Neutrino losses also have a minor influence in this fast evolution (Barkat et al. 1967), so their effect is included but negligible. The hydrodynamic evolutions are simulated using a one-dimensional PPM (piece-wise parabolic method) code (Colella & Woodward 1984). A small nuclear reaction network, which contains only 13 alpha-nuclei from 4 He to 56 Ni, is added to the hydrodynamics model to determine the rate of nuclear energy generation and the composition profiles. The detailed nuclear evolution is then calculated by a post-processing, using the thermodynamic trajectories from our hydrodynamic models and a large reaction network which contains 283 isotopes up to Br (Hix & Thielemann 1996). Models of 150, 170, 200 and 3OOM0 stars are calculated. Figure 1 shows the evolutionary paths of the central temperature and density. Table 1 sum­ marizes the physical quantities during the evolution and the amount of major elements in the ejected matter. The 170 M Q model becomes dynamically unstable at the central tempera­ ture Tc ~ 1.7 x 109 K and starts to collapse. In the early phase the dynamical evolution proceeds almost adiabatically with log T/logp ~ 1/3. The temper­ atures in the central region reach Tc ~ 3 x 109 K during the collapse and explosive oxygen burning sets in. The central temperature then rises above Tc = 3.5 x 109 K and heavy elements are synthesized. The inner region gradu-

364 10

i

•

n

'

i

'

r

Fe—decomposition (T< 4 / 3 )

,-9.5

Oxygen ignition

E-"

o

Onset of the dynamical instability

ao

9 4

5 logioPc

6 (g/cm3)

7

Figure 1: Schematic evolution of central density and temperature of a 170M© star. Two hatched areas show instability regions where T < 4 / 3 . The evolution of a 170M© star after the dynamical instability is presented.

ally exits the instability region and the nuclear energy release reverses the infall of the core into explosion. Finally the star undergoes complete disruption. Figure 2 displays the composition profile of the ejecta as functions of en­ closed mass and the expansion velocity (left) and the abundance ratios of stable isotopes relative to solar system values (right). Neutron-rich isotopes are less produced while the alpha-chain isotopes are more abundant than solar's. Also the intermediate nuclei from Si to Ca are synthesized very abundantly. In ad­ dition, odd-z isotopes, such as 23 Na, 27 A1, 3 1 P, 37C1 and 39 K, are less produced because of the use of metal-free stars as progenitors (see Umeda et al. 1999). The final explosion energy E = 1.5 x 1052 erg is much greater than that of a usual Type II supernova (for SN 1987A, E = (1.0 - 1.5) x 1051 erg; Shigeyama et al. 1990). The 200M Q star gives qualitatively the same result as that of the 170MQ star. However, the collapse proceeds to the higher central temperature, so iron is produced more abundantly and the explosion energy is greater than the 170MQ star. The calculation for the 3OOM0 star does not lead to the explosion. Most of the oxygen fuel is consumed and temperatures exceeds 7 x 109 K. Then

365 1600 2400

2700

2800

enclosed m a s s (M/MQ)

V (km/s)

Mass N u m b e r

Figure 2: Left: Final nucleosynthesis yields from a 17OJW0 Pop III pair-instability supernova. Only dominant abundances are presented. Right: Abundances of stable isotopes relative to the solar system values for a 170 M© star. The ratios are normalized to 1 6 0 . Isotopes of various elements are connected by solid lines.

the nuclei are photo-disintegrated into alpha-particles and nucleons, which consume more energy than previously released by oxygen and silicon burning. Thus, the star will collapse towards a black hole. Our lowest mass model of 150 M Q explodes as a pair-instability supernova. However, only 5 M© of oxygen is burned and the nuclear energy release from oxygen burning is relatively small. The final explosion energy E = 2.3 x 1051 erg is comparable with that of a usual Type II supernova (SN II). We also calculate theoretical light curves for pair-instability supernovae (Figure 3) with a one-dimensional spherically symmetric radiative transfer code (Iwamoto et al. 2000). We find that the light curve of a 170MQ pair-creation supernova is very bright (L ~ 10 43 erg s _ 1 ), about ten times that of usual SNe II, because of the large production of 56 Ni. The light curve keeps very luminous, even after it reaches the late tail phase around 300 days, in which the ejecta is thin against optical photons and the light curve is determined by the decay rate of 56 Co —> 56 Fe. It is interesting to compare with the brightest SN II 1997cy (Germany et al. 2000; Turatto et al. 2000), although the model curve declines faster than SN 1997cy. Our results are qualitatively consistent with previous studies (e.g., Herzig et al. 1990, Woosely k. Weaver 1982).

366 Table 1: Physical values of the pair-instability supernovae and their main nucleosynthesis products for all models.

Progenitor mass (MQ) He core mass (MQ) Tpeak (xlO 9 K) Ppeak ( X l O 6 g / c m 3 )

Explosion energy (xlO 51 erg) 4 He(M 0 ) 12 C(M 0 ) 16

24

44

Ti Cr 52 Fe 56 Ni 48

3

0(MQ)

Mg(M 0 ) 28 Si(M 0 ) 4O Ca(M 0 ) (decay into 44 Ca) (MQ) (decay into 48 Ti) (M©) (decay into 52 Cr) (M©) (decay into 56Fe) (M 0 ) Fe (MQ) Mg (MQ) [Fe/H] [Mg/H]

150 81 3.5 2.0 2.3 45 3.7 49 4.3 6.9 0.4 5.1 x 10" 5 1.7 x 10~4 1.8 x 10" 3 3.6 x 10" 2 5.8 x 10" 2 4.26 -3.58 -1.43

170 88 4.0 2.2 15 47 3.0 47 3.6 14 1.1 1.5 x 10" 4 2.5 x 10- 3 3.6 x 10" 2 0.69 0.74 3.6 -3.31 -2.33

200 102 4.5 3.5 17 66 2.7 53 4.1 16 1.3 1.8 x 10" 4 5.0 x 10" 3 1.0 x 10" 1 2.7 2.8 4.1 -2.80 -2.35

300 146 (collapsed)

Discussion Sz Conclusion

As we have seen in §2, pair-instability supernovae explode very energetically and produce large amounts of intermediate mass nuclei. Their abundance pat­ terns are quite different from those of usual Type II supernovae (Thielemann et al. 1996). However, models for the recently-discovered very energetic su­ pernovae, called hypernovae, such as SN1998bw and SN1997ef (Iwamoto et al. 1998, 2000; Nakamura et al. 2000), show similar abundance patterns as pairinstability supernovae. Both types of supernovae have large explosion energies and produce plenty of Si, S, Ar, Ca and Fe. The abundance data observed in metal-poor halo stars shows large starto-star variations and it suggests that the interstellar medium (ISM) of the halo was not well mixed and prior nucleosynthesis involved a single supernova or a few supernova (e.g., Audouze & Silk 1995; Shigeyama & Tsujimoto 1998). That is, the metal-deficient stars may have information on abundances of indi­ vidual supernovae which exploded at the early epochs of the Galactic evolution. We note that the observations of very metal-deficient stars (McWilliam et al.

367

iQ

4 0

I I

0

I

I

I

I

I

100

I I

1

1 I

200

I I

I

I

i

300 day

I

I

I

|

I

400

I

I

I

I

I

500

I

I

I I

600

Figure 3: A theoretical light curve for a 170M© population III pair-instability supernova with 6 6 Ni of O.69Af0 (solid line). The observed light curve of SN1997cy is also shown (filled circles; Turatto et al. 2000).

1995a,b; Ryan et al. 1996) have shown the interesting trends of the abun­ dances of iron-peak elements. The elements Mn and Cr become very underabundant in stars with lower metallicity ([Fe/H] < —2.5; [X/Y] ='logio(X/Y) — l°gio(X/Y)0), while Co becomes overabundant (Nakamura et al. 1999). Comparing with the observations, the abundances of a-elements in the 170 and 200 M Q pair-creation supernova ejecta are remarkably larger, especially the ratios [Si, S, Ar, and Ca/Fe]. On the other hand, [C/Fe] is smaller than the solar ratio. Theoretical abundances of iron-group nuclei are also different from those of metal-deficient stars and the above trend cannot be reproduced. [Cr/Fe] is larger than the observed abundance ratio while [Co/Fe] is much smaller (Note that too small [Co/Fe] is the common problem of SNe II yield; Nakamura et al. 1999). Although we have looked for the metal-deficient stars which have the same abundance patterns as our models, we have not found a good example yet. It suggests that pair-instability supernovae were not dominant in the early Galaxy. However, there still remains a possibility that a part of metal-poor stars was contaminated by the ejecta of pair-instability supernovae. The num­ ber of well-observed metal-poor stars is too small (~ 50) to conclude that no

368

effect of pair-instability supernovae is seen in the early Galaxy, taking into account the theoretical occurrence ratio of this type of supernovae to all the types (~ 1/300). The characteristic abundance patterns of pair-instability supernovae presented in this paper, such as high [Si, Ca/Mg (or Fe)], might be a clear evidence. Future observations with a large telescope such as SUBARU might make it possible to find any metal-deficient stars which have the abundance pattern of pair-instability supernovae ejecta. It will provide us im­ portant information on the existence of pair-instability supernovae and their effect on the Galactic chemical evolution. This work has been supported by the grant-in-Aid for Scientific Research (12640233) and COE research (07CE2002) of the Japanese Ministry of Educa­ tion, Science, and Culture. References 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16.

Audouze, J., & Silk, J. 1995, ApJ, 451, L49 Barkat, Z., Rakavy, G., Sack, N. 1967, Phys. Rev. Lett. 18, 379 Colella, P., & Woodward, P. R. 1984, J. Comput. Phys. 54, 174 Figer, D. F. et al. 1998, ApJ, 506, 384 Fowler, W. A., & Hoyle, F. 1964, ApJS. 9, 201 Fraley, G. S. 1968, Ap&SS., 2, 96 Germany, L.M., Reiss, D.J., Schmidt, B.P., Stubbs, C.W., Sadler, E.M. 2000, ApJ, 533, 320 Hix, W. R. & Thielemann, F.-K. 1996, ApJ, 460, 869 Herzig, K., El Eid, M. F., Fricke, K. J. & Langer, N. 1990, A&A, 233, 462 Hutchins, J. B. 1976, ApJ, 205, 103 Iwamoto, K. et al. 1998, Nature, 395, 672 Iwamoto, K. Nakamura, T., Nomoto, K., Mazzali, P. A., Garnavich, P., Kirshner, R., Jha, S., Balam, D. & Thorsternsen, J. 2000, ApJ, 534, 660 McWilliam, A., Preston, G.W., Sneden, C., & Shectman, S. 1995a, AJ, 109, 2736 McWilliam, A., Preston, G.W., Sneden, C., & Searle, L. 1995b, AJ, 109, 2757 Nakamura, T., Umeda, H., Nomoto, K., Thielemann, F.-K., & Burrows, A. 1999, ApJ, 517, 193 Nakamura, T., Maeda, K., Iwamoto, K., Suzuki, T., Nomoto, K,, Maz­ zali, P.A., Turatto, M., Danziger, I.J., & Patat, N. 2000, in IAU Symp. 195, Highly Energetic Physical Processes and Mechanisms for Emission

369

17. 18. 19. 20. 21. 22.

23. 24. 25. 26. 27.

from Astrophysical Plasmas, eds. P. Martens, S. Tsuruta, k M. Weber (PASP), p.347 Nomoto, K. k Hashimoto, M. 1988, Phys. Rev. Lett. 163, 13 Shigeyama, T., k Nomoto, K., 1990, ApJ, 360, 242 Shigeyama, T., k Tsujimoto, T. 1998, ApJ, 507, L135 Slattery, W. L., Doolen G. D., k De Witt H.E. 1982, Phys. Rev. A, 26,2255 Thielemann, F.-K., Nomoto, K., k Hashimoto, M. 1996, ApJ, 460, 408 Turatto, M., Suzuki, T., Mazzali, P.A., Benetti, S., Cappellaro, E., Nomoto, K., Nakamura, T., Young, T.R., Patat, F. 2000, ApJ, 534, 57 Ryan, S. G., Norris, J. E., k Beers, T. C. 1996, ApJ, 471, 254 Tohline, J. E. 1980, ApJ, 239, 417 Umeda, H., Nomoto, K., Nakamura, T. 2000, in The First Stars, eds.: Weiss, Abel, Hill, in press (astro-ph/9912248) Umeda, H., k Nomoto, K., 2000, in preparation Woosley, S. E.,& Weaver, T. A. 1982, in Supernovae: A Survey of Current Research, eds. M. J. Rees and R. J. Stoneham, Dordrecht: Reidel, p.79

X - r a y O b s e r v a t i o n s of S N R s a n d h o t I S M i n t h e Large M a g e l l a n i c Cloud —the c h e m i c a l e n r i c h m e n t of t h e g a l a x y

Mamiko Nishiuchi, Jun Yokogawa, Ichizo Hayashi, Katsuji Koyama Department of Physics, Faculty of Science, Kyoto University Kitashirakawa-Oiwake-cho, Sakyo-ku, Kyoto, 606-8502, Japan E-mail: [email protected] John P. Hughes Department of Physics and Astronomy, Rutgers University 136 Frelinghuysen Road, Piscataway, NI 08854-8019 The great portion of the elements are thought to be produced in'the supernova (SN) explosions and gradually mixed into the interstellar matter (ISM) of the galaxies. We can trace the course of the chemical-pollution of the galaxies by observing the supernova remnants (SNRs) and ISM. We present the X-ray measurements of metal abundances of the Large Magellanic Cloud (LMC). All the archive data in the vicinity of the LMC taken with the Advanced Satellite for Cosmology and Astro­ physics (ASCA) were used. The X-ray spectroscopy of the diffuse X-ray emission spreading over a large portion of the LMC was carried out in order to measure the metal abundance of the ISM directly. With the good spectral resolution of ASCA the nature of this diffuse X-ray emission was first confirmed to be thermal in origin, likely to be emitted from hot ionized ISM in the galaxy, because ASCA detected line emissions from various elements, which was beyond the capability of previous X-ray satellites. The X-ray spectrum of diffuse X-ray emission was reproduced by the Non-Equilibrium Ionization (NEI) model with temperature of ~ 1.2 keV. The overall elemental abundances determined from X-ray spectroscopy of ISM is found to be also consistent with previous results determined through X-ray SNR spectroscopy derived by Hughes, Hayashi and Koyama (1998) and previous optical and UV analysis, except overabundance of sulfer.

1

Introduction

SNe and their remnants are the key to our understanding of the origin of the chemical elements in the universe. The nucleosynthesis process inside a star generates many elements and SN explosions blow t h e m out into the interstellar space. Finally blown out elements are thought to be gradually mixed into the ISM uniformly. In the hot plasmas (10 6 ~ 10 7 K) both of the SNRs and the ISM of the galaxies, most of the elements are ionized up to the H-like or He-like ion stages, emitting strong line emissions in the X-ray band. Therefore, the X-ray spectroscopy of the SNRs and the hot ISM in the galaxies can tell us much information, for example, what kind of and how much elements are ejected by

370

371

the SN explosions (from the X-ray spectra of young aged SNRs), are contained in the ISM (from the X-ray spectrum of middle aged SNRs and ISM itself), and so on. This information leads us to the understanding of the chemical evolution of the galaxies. Because of the proximity, reasonable angular sizes (not too large) and the low interstellar absorption toward it, we selected the Large Magellanic Cloud (LMC), known as the satellite galaxy of our own, as a target to investigate this problem using the X-ray spectroscopy. Because the LMC is a younger galaxy than our own, investigating it means to probe the aspect of our Galaxy at early stage of its life. The LMC is known to have many X-ray emitting SNRs inside. From the systematic X-ray spectral analysis of middle aged SNRs (Hayashi PhD thesis 1997 and Hughes, Hayashi and Koyama 1998), the abundances of the elements in the ISM around SNRs are obtained. This method was completely independent of previous ones to get the abundances of elements, such as UV and optical observations. The derived overall abundance pattern was consistent with previous results. The LMC, at the same time, is known to have plenty of diffuse X-ray emis­ sion. Since satellites with imaging capability came into use, such as Einstein and ROSAT, many observations of this diffuse X-ray emission were carried out. However, the spectral resolution of both the Einstein and ROSAT was not good enough to know the nature of the diffuse X-ray emission, namely could not tell thermal from non-thermal in origin. In the following we present the abundance determination through the ASCA X-ray spectroscopy of the hot plasma in the LMC. In section 2, we briefly introduce the abundance determination using the X-ray spectroscopy of SNRs carried out by Hughes, Hayashi and Koyama (1998). Section 3 provides the description of the observations and results of the diffuse X-ray emission in the galaxy. Finally we discuss the abundance pattern derived from these two methods. 2

X-ray Spectroscopy of S N R s

Evolution of SNRs are considered to be classified into three phases; 1: The initial phase of evolution is so-called "free-expansion phase". Almost all the initial explosion energy transit into the kinetic energy of the ejected matter. The ejected matter is heated up at the shock front. In this phase the observed X-ray spectrum has strong line emissions of various elements contained in the ejecta of SN explosions. 2:The second phase is known to be "adiabatic phase" or "Sedov phase". In this phase the shell is decelerated and at the same

372

time significant amount of ISM is swept up. So the X-ray spectrum traces the characteristics of heated up ambient ISM. 3: The final stage of the SNR evolution is called "radiative-cooling phase". In this stage, the radiative cooling begins in earnest. The cold and dense shell, which has no longer capability to emit X-ray emission, is left. Therefore, one can use the X-ray spectroscopy of SNRs in the adiabatic phase as a tool to determine the chemical composition of ISM. Hughes, Hayashi and Koyama (1998) made a systematic X-ray spectroscopy of intermediate age SNRs in the LMC, which are thought to be in the adia­ batic phase, using the ASCA data. Because of the poor angular resolution of ASCA, the SNRs at the distance of the LMC could not be spatial resolved. Therefore, in the spectral fitting analysis, they used the "Sedov self-similar solution model"; they divided the X-ray emitting region within the SNR into many shells in order to consider the radial distribution of density of the elec­ trons and ions and temperature of plasma inside. The entire spectrum of SNR is composed of those emitted by the series of shells. The derived abundances are shown in figure 1. The overall abundances are consistent with the previous results obtained by the optical data, but do not show anomalous overabundance of magnesium and silicon seen by the Russell & Dopita (1992). 3

X-ray Spectroscopy of ISM

ASCA had observed the LMC 53 times by the end of Sep. 1999. We used all those in the archive (49 pointing observations) and those taken in the LMC survey project during the ASCA Announce of Opportunity (AO) 7 phase (9 pointing observations), which are our own proposed observations. The total exposure time is 1920 ksec. All the archive data were obtained from NASA HEASARC (High Energy Astrophysics Science Archive Research Center) data system. Most of the optically bright position called optical bar and the region towards the northern part of the galaxy were covered with these data. After removing all the discrete sources which exceeded over 5 sigma signalto-noise level threshold, we accumulated the diffuse X-ray spectra from each pointing observation around optical bar and summed them up to obtain good statistics. In the spectral fitting, we tested the non-thermal model. However this model was completely rejected not only from the statistical point of view but also from the waving behavior of residuals from the best fit model, indicating the existence of the line emissions from various elements. On the contrary the thermal plasma models in which elemental abundances were treated as free pa-

373

rameters were more appropriate for both of Collisional lonization Equilibrium (CIE) and Non-Equilibrium lonization (NEI) state conditions. The tempera­ ture was determined to be ~ 0.7 keV for CIE and ~ 1.2 keV for NEI model, respectively. This result shows, for the first time, that the diffuse X-ray emis­ sion in the LMC is thermal in origin, probably coming from the hot ISM of the LMC. At the same time, the NEI plasma model was more favorable than CIE one because the CIE plasma model required unreasonable high abundances of sulfur, argon and calcium, though the same kind of tendency was also seen, still less, in the NEI model. The derived abundances from NEI model fitting of hot ISM spectrum in the LMC are also shown in figure 1. The overall pattern of abundances derived from one temperature plasma model fitting are consis­ tent with the previous ones. However, the elements whose line energies are higher than sulfur take higher abundances and at the same time the elements whose line energies are lower than that of sulfur is in a little insufficient to the abundances obtained from the previous analysis. This condition maybe results from the artifact of the one temperature plasma model fitting and suggests us an implication of other higher temperature plasma component. Our future work will be devoted to the more detail abundance determination of the hot ISM with two temperature thermal plasma models.

Mean LMC Abundances

Figure 1: Mean LMC abundances derived from optical/UV observations (Russell &i Dopita 1992; RD92) and the X-ray spectroscopy of middle-aged SNRs (Hughes, Hayashi and Koyama 1998) and ISM (this work)

References 1. Russell, S.C., & Dopita, M.A., 1992, ApJ, 384, 508 2. Hayashi, I. PhD thesis, 1997, ISAS Research Note 620. 3. Hughes, J.P., Hayashi, I., and Koyama, K., 1998, ApJ, 505, 732

R-PROCESS NUCLEOSYNTHESIS IN NEUTRINO-DRIVEN W I N D : GENERAL RELATIVISTIC EFFECTS A N D SHORT DYNAMIC TIMESCALE MODEL

KAORI OTSUKI, TOSHITAKA KAJINO Division

of Theoretical

Astrophysics, National Astronomical Observatory Mitaka, Tokyol 81 -8588, Japan E-mail: [email protected], [email protected]

of

Japan,

SHIN-YA W A N A J O Department of Physics, Sophia University, 7-1 Kioi-cho Chiyoda-ku Tokyo 102-8554, Japan E-mail: [email protected] HIDEYUKI TAGOSHI Graduate

Department of Earth and Space Science, School of Science Osaka University, Toyonaka OSAKA E-mail: [email protected]

560-0043,

Japan

Neutrino-driven wind from young hot neutron star, which is formed by supernova explosion, is the most promising candidate site for r-process nucleosynthesis. We use spherical steady state flow model in general relativistic framework. Exploring wide parameter region which determines the expansion dynamics of the wind, we can find interesting physical conditions which lead to successful r-process nucle­ osynthesis.

1

Introduction

Neutrino-driven wind is one of the most promising candidates for r-process nucleosynthesis. It is generally believed that a neutron star is formed as a remnant of gravitational core collapse of Type II, lb or Ic supernovae. The hot protoneutron star releases most of its energy as neutrinos during KelvinHelmholtz cooling phase, and these neutrinos drive matter outflow from the surface. This outflow is called neutrino-driven wind. We briefly explain a r-process nucleosynthesis scenario in neutrino-driven wind. On the surface of a protoneutron star (the temperature T ~ 3-5 MeV), the ejected material consists of almost free neutron and proton. They are blown away from the surface and begin to cool down. When the temperature reaches T ~ 0.5 MeV, most protons are assembled into a-particles and acapture process begins. This is a process to produce seed nuclei. If the temperature subsequently reaches T ~ 0.3-0.2 MeV, charged particle reactions

374

375

begin to freeze out and are followed by r-process nucleosynthesis. This scenario is consistent with observations which suggest that the r-process should be primary process 7 . Nevertheless, various numerical calculations of neutrino-driven winds do not agree on the entropy and show essentially different results of r-process nucleosynthesis calculation in the wind 2 6 8 9 10 . We use spherical steady state flow model in order to avoid uncertainty arising from numerical simulation methods, supernovae models, etc. We investigate suitable physical conditions for r-process nucleosynthesis in neutrino-driven wind. 2

Model

Since the wind blows near the surface of the neutron star, it is needed to study expansion dynamics of neutrino-driven wind in general relativity. Although there might be some conditions, such as rotation and convection, which break the spherical symmetry and steady states, we here assume the spherical sym­ metric and steady state cases in order to investigate the essential physical properties of neutrino-driven wind. The basic equations to describe the spherically symmetric and steady state winds in Schwarzschild geometry are given by M = 4Trr2pbu, du 1 dP ( , u— = — — 1+u2 dr ptot + P dr de dr

P dpb pi dr

(1) 2M\

M -—,

(2)

(3)

where M is the mass outflow rate, r is the distance from the center of the neutron star, pb is the baryon mass density, u is the radial component of the four velocity, ptot = Pb + Pb£ is the total energy density, e is the specific internal energy, P is the pressure, M is the mass of the neutron star, and q is the net heating rate due to weak interactions 5 6 . This heating rate depends on neutrino luminosity. We use the conventional units that the plank constant h, the speed of light c, Boltzmann constant fc, and gravitational constant G, are taken to be unity. We assume that the material in the wind consists of photons, relativistic electrons and positrons, and non-relativistic free nucleons. We fix T=0.1 Mev at 104 km to define the mass loss rate. In this calculation, we adopt the typical values of a protoneutron star: the radius r; = 10 km,and the baryon density at the surface pi = 10 10 g/cm 3 . Solving these differential

376

Figure 1. Relations of S and T^yn.

equations, we can obtain radial profiles of velocity, temperature, and density for various parameter sets of neutron star mass M and neutrino luminosity

3

Results

One of the most important hydrodynamic quantity, which characterizes the expansion dynamics of the neutrino-driven wind, is the duration time of the a-process; called dynamical timescale Tdyn. It is defined as a timescale that the temperature decreases one e-fold from 0.5MeV. Entropy per baryon, S, settles the fraction of free nucleon in NSE material. The electron fraction gives a ratio of neutron and proton in material. Since r-process nucleosynthesis is very sensitive to the electron fraction, it is regarded as free parameter in this work. Now we can calculate S - Tdyn relations for various neutron star masses and neutrino luminosities. We show the results in Fig. 1. Two hatched regions satisfy the approximate condition which produce A=200,130 nuclei 3 5 . We find that dynamical timescale as short as Tdyn — 6 ms with M = 2.0 M Q and Lv = 1052 ergs/s is preferable in this figure. Note that, in Newtonian framework, we can not find such a short dynamic timescale wind in 6 . Finally, we confirm that the second and third peaks are produced with these conditions by network calculations. Figure 2 shows final r-process abun­ dances in this calculation compared with these of the solar system 4 . The

377

LogY

0

20

40

60

80

100

MASS

120

140

160

180

200

220

240

NUMBER

Figure 2. r-process abundance in our calculation and solar system.

r-process abundances in the solar system are adopted in arbitrary unit. For successful r-process nucleosynthesis, we can find interesting physical conditions, short dynamical timescale Tdyn ~ 6 ms and relatively low entropy S ~ 140. Unfortunately, we could not find such conditions on M = 1.4M0 case so far as spherical steady state flow. A continued effort in investigating various effects on neutrino-driven winds with detailed calculations would be required to reproduce r-process nucleosynthesis in the solar system. References 1. Burbidge, E.M., Burbidge, G.,R., Fowler, W.A., & Hoyle,F. Rev.Mod.Phys. 29, 547 (1957) 2. Cardall, C.Y. & Fuller, G. ApJL 486, L l l l (1997) 3. Hoffman, R. D., Woosley, S. E.,& Qian, Y.-Z. ApJ 482, 951 (1997) 4. Kappeler, F., Beer, H., & Wisshak, K. Rep.Prog.Phys. 52, 945 (1989) 5. Otsuki, K., Tagoshi, H., Kajino, T., & Wanajo, S. ApJ 533, 424 (2000) 6. Qian,Y.-Z., & Woosley, S., E. 1996 471, 331 (1996) 7. Sneden, C , McWilliam, A., Preston, G., Cowan, J.J.,Burris, D.L., & Armoski, B.J. ApJ 467, 819 (1996) 8. Takahashi, K., Witti, J., & Janka, H.-Th. A&A 286, 857 (1994) 9. Witti, J., Janka, H.-Th., & Takahashi, K. A&A 286, 842 (1994) 10. Woosley, S.E., Wilson,J.R., Mathews,G.J., Hoffman, R.D., & Meyer, B.S. ApJ 433, 229 (1994)

The Critical Role of Light Neutron-Rich Nuclei in the r-Process Nucleosynthesis Mariko Terasawaa, K. Sumiyoshib, T. Kajinoc, I. Tanihata d , G. J. Mathews6 and K. Langankef a

The University of Tokyo, Bunkyo, Tokyo 113-0033, Japan

b

Numazu College of Technology (NCT), Numazu, Shizuoka 410-8501, Japan

c

National Astronomical Observatory (NAO), Mitaka, Tokyo 181-8588, Japan

d

The Institute of Physical and Chemical Research (RIKEN), Wako, Saitama 351-0198, Japan e

The University of Notre Dame, Notre Dame IN 46556

f

Institute of Physics and Astronomy, University of Aarhus, Denmark

We have extended the nuclear reaction network for 1 < Z < 10 in order to investigate the role of light neutron-rich nuclei in the r-process. We find that a new nuclear reaction flow path opens in the region of very light neutron-rich nuclei as the temperature decreases. This'new reaction flow can affect the final abundances by up to a few orders of magnitude, while still producing the characteristic three r-process peaks as well as the hill of the rareearth elements.

1. Introduction Most previous studies of r-process nucleosynthesis have been largely concerned with the reaction flow through heavy unstable nuclei. Typically, nuclear reaction networks have included a few thousand heavy nuclei, while only a limited number of the light-mass nuclei near the valley of stability were considered. However, if the r-process occurs in i/-driven winds material in the wind expands from a high-entropy hot plasma consisting initially of free neutrons, protons, and electron-positron pairs under an intense flux of neutrinos. In such conditions, neutron-rich light-mass nuclei as well as heavy nuclei can play an important role in the production of both seed nuclei and r-process elements. We therefore have extended the nuclear reaction network to include ~ 80 unstable nuclei up to the neutron-drip line for 1 < Z < 10. We compare the final r-process abundances in the i/-driven winds with and without this extension of the network.

378

379 2. Nuclear Reaction Network For 10 < Z < 94, we have used the network of Meyer et al. [1], which includes about 3000 nuclear species from the /^-stability line to the neutron drip line. However, for our purpose, their network is too narrow for light nuclei with Z < 10. It includes only ten nuclei n, p, 2H, 3H, a, 9Be, 12 C, 13C, 1 6 0, 1 7 0. We have extended this network to include 80 nuclei up to the neutron-drip line for Z < 10 (see dots in Fig. 1). Our network (hereafter called 'the full network') includes almost all charged-particle reactions for A< 28. For comparison, we have also used the narrower network ('the small network'). For the ^-interactions, we have included ve capture for all nuclei, as well as ve capture by free protons ([2]). For very neutron-rich nuclei, we have also included ^-induced neutron emission([3]) (see Terasawa et al. [4] for more details). 3. ^-driven winds We employ the ^-driven wind models of Sumiyoshi et al. [5]. We choose the case in which the expansion timescale is as short as Tdyn = 0.005 sec with M^ 2.0 M 0 , RNS = 10 km and, L„itotai = 6 x 1052 erg/sec. For illustration, we use one typical trajectory of the wind. We start the r-process network calculation from the time when the temperature drops to T 9 = 9.0. At this point, the initial nuclear statistical equilibrium consists of free neutrons and protons. The initial Ye 0.42 is taken from the hydrodynamical simulations. 3.1. Results

m-3

20

1 1 1

i

n . i

( 1

,

. I ,

,

i

, ,

| .

si,/

8 0 fig) ■ ® ■ ® 0 « 5|B©c3

O

5

5

10

i

i

15 N

D

Ys= 10-' 10- t0-'° i

,

20

ri

C

5

gt-"

0

O

Abundance

e s ® e> G ^ | B ® 0

N 10

O

t - 0.567 (sec)

15

25

30

Figure 1. The abundances and a main path for Z< 15.

0

wp 50

N i y■ !'■■■ '■-

100 150 200 Mass Number

l.,I.A...L.J.
'. However, previ­ ous works were performed by using specific nuclear models and values of these quantities are designated almost uniquely in each previous works (e.g. 3 ' 4 etc.).

381

382 .

I

1

,

■

,

I

1

I

I

■

i

i

i

i

i

.

0^400: cylinder \

\

\

^ \ \

"

, sphere

;

0.05

uniform

\

1 slab \ \ \ \ spherical . A/ hole

■ 0.1

f

(^cylindrical "T hole

, 0.15

0.2

3

n„ (fm- ) Figure 1: Zero-temperature phase diagram on the n^ (baryon number density) versus C2 (its standard value is 1.0) plane, evaluated for C\ = 400 MeV fin2. (400 MeV fm2 is its standard value.)

2

Results

In the present work, by generalizing a compressible liquid-drop model devel­ oped by Baym et al.6 in such a way as to incorporate uncertainties in Esur( and (o) Here we set them as MP

/#> = -CxnT Esur{ = C 2 tanh I - ^ y

(1) (2)

where Ci, C2 and C 3 are the adjustable parameters being positive definite; fj,ii is the neutron chemical potential in the neutron gas; E^f is the surface tension used by BBP. Eq. (1) approximately reproduces the overall density dependence of the results obtained from the Hartree-Fock theory with Skyrme interactions and from the lowest-order Brueckner theory with the Reid soft-core potential. As for Eq. (2), if we set C2 = 1.0 and C3 = 3.5 MeV (hereafter, C 3 is fixed at 3.5 MeV), .Esurf agrees well with the RBP's Hartree-Fock calculation7. The resultant phase diagram (Fig. 1) show that while the phases with cylindrical holes and with spherical holes can exist only for unrealistically low •E'surf, the phases with cylindrical nuclei and with slab-like nuclei survive almost

383

cylinder

a

0.01 r

/iS-

Figure 2: The critical temperature Tc for the phases with planar and cylindrical nuclei as a function of baryon density rib- The thick curves lying between the two vertical lines are the results in the density region in which the phase with planar (or cylindrical) nuclei is energetically stable.

independently of J5surf and /J,\, '. For these two phases, as noted by Pethick and Potekhin5, elastic properties of the nuclear rods and slabs are characterized by elastic constants used for the corresponding liquid-crystal phases. We also estimate thermally induced displacements of these nuclei using their expressions. In Fig./ 2, we plot the critical temperature at which the relative displacements y/{\v\2)/(a/2 — TN), which mean the shortest distance between the surface of the nucleus in its equilibrium position and the boundary of the cell containing it, become unity. Focusing on the typical values of C\ = 400 MeV fm2 and C2 = 1.0, we can see that the layered lattice of slab-like nuclei will be melt at typical temperatures of the neutron star crusts (&BT ~ 0.1 MeV) by thermal fluctuations. The difference in T c between these two lattices, as can be observed from Fig. 2, suggests that if formation of these two lattices can occur dynamically in the star, the layered phase is formed later than the triangular phase during the star' s cooling. It is expected that latent heat is released when these phases are formed. We also estimated the amount of released latent heat Q, and found that it yields Q ~ 1046 erg. Its effect seems to be detectable if it occurs after the neutron star cools down enough.

384

3

Conclusion

We have examined the dependence on the surface tension ESUI{ and on the proton chemical potential fj,p in pure neutron matter, of the density region in which the presence of non-spherical nuclei and of bubbles is energetically favored at T = 0. We have found that as ESUT{ decreases or /xp ' increases, such a density region becomes larger. For the values of Eami and fj,p ' as adopted in recent literature, our results show that in the ground state, the phases with rod-like nuclei and with slab-like nuclei lie between the bcc lattice phase and the uniform nuclear matter phase. The fluctuational displacements of such non-spherical nuclei from their equilibrium positions have been estimated at finite temperature. It has been suggested that at temperatures typical of matter in the neutron star crust, such fluctuations may melt the layered lattice of slab-like nuclei. If the dynamical processes leading to the formation of the non-spherical nuclei, latent heat would be released when they occur. It is found that its effect is expected to be detectable if it occurs in old neutron stars. Acknowledgements We are grateful to Professor Takeo Izuyama for useful discussion and valuable comments. References 1. D.G. Ravenhall, C.J. Pethick and J.R. Wilson, Phys. Rev. Lett. 50 (1983) 2066. 2. M. Hashimoto, H. Seki and M. Yamada, Prog. Theor. Phys. 71 (1984) 320. 3. C.P. Lorenz, D.G. Ravenhall and C.J. Pethick, Phys. Rev. Lett. 70 (1993) 379. 4. K. Oyamatsu, Nucl. Phys. A561 (1993) 431. 5. C.J. Pethick and A.Y. Potekhin, Phys. Lett. B427 (1998) 7. 6. G. Baym, H A . Bethe, C.J. Pethick, Nucl. Phys. A175 (1971) 225 (BBP). 7. D.G. Ravenhall, C D . Bennett and C.J. Pethick, Phys. Rev. Lett. 28 (1972) 978 (RBP). 8. G. Watanabe, K. Iida and K. Sato, astro-ph/0001273.

Symposium Program

This page is intentionally left blank

Symposium Program

The Origin of Matter and Evolution of Galaxies 2000 January 19 (Wednesday) Open i ng M. Ishihara (Director of Center for Nuclear Study, U. Tokyo) S. Kubono (Co-Chairperson, CNS) Early Universe and Chemo-Dynamic Evolution of Galaxies I G. Mathews (U.Notre Dame) , Chair T. Kajino (NAO) Prospect of Nuclear Cosmology; From Big-Bang to Supernovae T. Shigeyama (U. Tokyo) Inhomogeneous Chemical Evolution in the Galactic Halo T. Suzuki (NAO) A New Model of Evolution of Light Elements in Inhomogeneous Galactic Halo Early Universe and Chemo-Dynamic Evolution of Galaxies II S. Kubono (CNS), Chair T. Kifune (ICRR, U. Tokyo) Prospect of Very High Energy Gamma Ray Astronomy with Next Generation Imaging Cerenkov Telescope S. Yanagita (Ibaraki U.) Some Evidence for Supernova Origin of Galactic Cosmic Rays Y. Ishimaru (U. Tokyo) The First Enrichment of the Galaxy Inferred from Neutron-Capture Elements H. Umeda (U. Tokyo) Nucleosynthesis in Massive Metal Free Stars Observation of Elements I T. Kajino (NAO), Chair S. Amari (Washington U.) Presolar Grains as Probe to Nucleosynthesis in Stars and Evolution of the Galaxy C. Iliadis (U. North Carolina) Nucleosynthesis in Globular Cluster Stars M.Y. Fujimoto (Hokkaido U.) Internal and External Process of Chemical-Pollution in Extremely Metal-Poor Stars Y. Fukazawa (U. Tokyo) X-Ray Measurements of Metal Abundances of Hot Gas in Clusters of Galaxies

387

388 Observation of Elements II M. Smith (ORNL), Chair H. Murakami (Kyoto U.) X-Ray Diagnosis of the Galactic Center Abundance with an "X-Ray Reflection Nebula" M. Teshima (ICRR) The Highest Energy Cosmic Rays N. Hasebe (Waseda U.) Cosmic Ray Observation for Nuclear Astrophysics; CORONA Program T. Kishimoto (Osaka U.) Kaonic Nuclei Excited by the (K~, N) Reaction Poster Session I (with 6 Minutes Presentation - No Question and Comment) M. Smith (ORNL), Chair A. Bamba (Kyoto U.) Chemical composition and Distribution of Heavy Elements in a Supernova Remnant H. Ishiyama (KEK) Tanashi Recoil Mass Separator for Nuclear Astrophysics A. Iwazaki (Nishogakusha U.) Col 1 ision between Neutron Star and Axion Star as a Possible Source of Gamma Ray Burst and Ultra High energy Cosmic Ray N. Kudomi (RCNP) Double Beta Decays of 100Mo by ELEGANT V at Oto Cosmo Observatory K. Maeda (U. Tokyo) Nucleosynthesis in Gamma-Ray Bursts and Abundances in Black Hole Binaries T. Minemura (Rikkyo U.) Coulomb Dissociation of 12N and 130 S. Nagataki (U. Tokyo) Effects of Jet-Like Explosion in SN 1987A N. Nakasato (U. Tokyo) Formation and Chemical Dynamics of the Galaxies January 19 (Thursday) Stellar Evolution and the Nucleosynthesis - Hydrostatic Burning J. D'Auria (TRIUMF), Chair R. Tribble (Texas A&M) Direct Capture S-factors from Asymptotic Normalization Coefficients N. Iwasa (RIKEN) Coulomb Dissociation of 8B for 7Be(p, y) N. Kudomi (RCNP) Beam Current Monitor and Gas Pressure Control Techniques of Development for 3He+3He Solar Reaction Nucleosynthesis in Explosive Burning and New Approach

389 R. Tribble (Texas M M ) , Chair M. Smith (ORNL) Probing Stellar Explosions with Radioactive Beams at ORNL J. D'Auria (TRIUMF) The DRAGON Facility for Nuclear Astrophysics Studies at the New ISAC Radioactive Beams Facility S. Kubono (CNS) New Radio-Isotope Beam Facility for Nuclear Astrophysics - Study of a Critical Stellar Reaction 150(a, y)19Ne T. Motobayashi (Rikkyo U.) Methods for Astrophysics Studies with Intermediate-Energy RI Beams Explosion of Massive Stars I R. Tribble, Chair S. Yamada (U. Tokyo) Physics of Collapse-Driven Supernovae M. Yasuhira (Kyoto U.) Protoneutron Stars with Kaon Condensate and Possibility of Delayed Collapse Explosion of Massive Stars II T. Motobayashi (Rikkyo U.), Chair R. Boyd (Ohio State U.) OMNIS, the Observatory for Multi-Flavor Neutrinos from Supernovae Y. Fukuda (ICRR, U. Tokyo) Observation of Supernova Neutrino at SuperKamiokande K. Homma (Hiroshima U.) Can the Negative Mass Square of the Electron Neutrino be an Indication of Interactions with Relic Neutrinos ? Poster Session II (with 6 Minutes Presentation - No Question and Comment) T. Motobayashi (Rikkyo U.), Chair J. Nakatsuru(U. Tokyo) Explosive Nucleosynthesis in Pair-Instability Supernovae M. Nishiuchi (Kyoto U.) X-Ray Observations of SNRs and Hot ISM in the Large Magellanic Cloud - The Chemical Enrichment of the Galaxy K. Otsuki (NAO) r-Process Nucleosynthesis in Neutrino-Driven Wind- General Relativistic Effects and Short Dynamic Timescale Model M. Serata (Rikkyo U.) Coulomb Dissociation of 13N and 140 M. Terasawa (NAO) New Nuclear React ion Flow towards r-Process Nucleosynthesis in Supernovae: A Critical Role of Light Neutron-Rich Nuclei 1