ASTROPHYSICS / ASTROPHYSIQUE

Transcription

Vol. 64 No. 4
OCTOBER-DECEMBER (FALL) 2008
OCTOBRE À DÉCEMBRE (AUTOMNE) 2008
Physics in Canada
La Physique au Canada
ASTROPHYSICS /
ASTROPHYSIQUE
Guest Editor / Rédacteur honoraire : Laurent Drissen, U. Laval
Serving the Canadian
physics community
since 1945 /
Au service de la
communauté canadienne
de physique depuis 1945
Canadian Association
of Physicists /
Association canadienne
des physiciens et
physiciennes
www.cap.ca
PHYSICS IN CANADA
LA PHYSIQUE AU CANADA
Canadian Association
of Physicists
Association canadienne des
physiciens et physiciennes
www.cap.ca
Vol. 64 No. 4 (October-December (Fall) 2008 / octobre à décembre (automne) 2008)
DE FOND
ARTICLES
DEPARTMENTS EDUCATION
DÉPARTEMENTS ÉDUCATION
FEATURES
199 Foreword - “Astrophysics”, by L. Drissen, Guest Editor
200 Préface - “Astrophysique”, par L. Drissen, rédacteur honoraire
201 L’Irradiance solaire et ses variations, by P. Charbonneau, A. Crouch et K. Tapping
207 Resonance Dynamics in the Kuiper Belt, by B. Gladman and J.J. Kavelaars
215 Visualizing the Invisible using Polarisation Observations, by J. Brown, J. Stil, and
J. Landecker
227 Metal-Poor Stars: The Intersection of Chemistry, Cosmology, and Stars,
by K. Venn
233
245
251
257
L’Evolution chimique des galaxies, by H. Martel
Results from the Gemini Deep Deep Survey, by R.G. Abraham et al.
Les galaxies à sursauts de formation stellaire dans l’ultraviolet, by C. Robert
Bringing Chemists and Physicists Together: The Legacy of the Ontario
Photonics Consortium at the University of Western Ontario, by R. Lipson
265 Women Physicists in Canada, by A. Predoi-Cross et al.
267 Poster presented by the Canadian Delegation to the 3rd IUPAP International
Conference on Women in Physics
268 International, Physics Olympiad 2008
by A. Kotlicki and N. Krasnopolskaia
226 Competition for a new PiC-PaC
logo / Concours pour un nouveau
logo PiC-PaC
250 Departmental, Sustaining, Corporate,
and Institutional Members / Membres
départementaux, de soutien, corporatifs, et institutionnels
Advertising Rates and Specifications (effective January 2008) can be found on the PiC website
(www.cap.ca - Physics in Canada).
Les tarifs publicitaires et dimensions (en vigueur depuis janvier 2008) se trouvent sur le site internet de
La Physique au Canada (www.cap.ca - La Physique au Canada).
Cover / Couverture :
Density of gas for a simulated
galaxy during the formation
of the thick disk – front view.
A galaxy is made of two
disks, one thick, the other
thin, enmeshed into one
another.
(Taken
from
Figure 3, page 236, of the
article by H. Martel)
Densité de gaz pour une galaxie
simulée durant l’époque de la
formation du disque épais – vue
de face. Une galaxie comprend
deux disques, l’un épais, l’autre
mince imbriqués l’un dans
l’autre. (Pris de la figure 3, page
236, de l’article par H. Martel)
LA PHYSIQUE AU CANADA / Vol. 64, No. 4 ( oct. à déc. (automne) 2008 ) C i
DEPARTMENTS
DÉPARTEMENTS
263 In Memoriam : Martin Wesley Johns
264 CAP News : 3rd IUPAP International
Conference on Women in Physics
270 2009 CAP Congress / Congrès 2009 de
l’ACP
PHYSICS IN CANADA
LA PHYSIQUE AU CANADA
The Journal of the Canadian Association of
Physicists
La revue de l'Association canadienne des physiciens et physiciennes
ISSN 0031-9147
EDITORIAL BOARD / COMITÉ DE RÉDACTION
Editor / Rédacteur en chef
272 Best Student Presentation Competition
at the 2009 CAP Congress / Compétition
pour la meilleure présentation étudiante
au Congrès 2009 de l’ACP
273 Books Received / Livres reçus
274 Book Reviews / Critiques de livres
276 Employment Ads / Affichage de postes
Béla Joós, PPhys
Physics Department, University of Ottawa
150 Louis Pasteur Avenue
Ottawa, Ontario K1N 6N5
(613) 562-5758; Fax:(613) 562-5190
e-mail: [email protected]
Associate Editor / Rédactrice associée
Managing / Administration
Francine M. Ford
c/o CAP/ACP; E-mail: [email protected]
Book Review Editor / Rédacteur à la critique de livres
Richard Hodgson, PPhys
c/o CAP / ACP
Suite.Bur. 112, Imm. McDonald Bldg., Univ. of / d' Ottawa,
150 Louis Pasteur, Ottawa, Ontario K1N 6N5
Email: [email protected]
Advertising Manager / Directeur de la publicité
Greg Schinn
EXFO Electro-Optical Engineering Inc.
400 av. Godin
Quebec (QC) G1M 2K2
(418) 683-0913 ext. 3230
e-mail: [email protected]
Board Members / Membres du comité :
René Roy, phys
Département de physique, de génie physique et d’optique
Université Laval
Cité Universitaire, Québec G1K 7P4
(418) 656-2655; Fax: (418) 656-2040
David J. Lockwood, PPhys
Institute for Microstructural Sciences
National Research Council (M-36)
Montreal Rd., Ottawa, Ontario K1A 0R6
(613) 993-9614; Fax: (613) 993-6486
Tapash Chakraborty
Canada Research Chair Professor, Dept. of Physics and Astronomy
University of Manitoba, 223 Allen Building
Winnipeg, Manitoba R3T 2N2
(204) 474-7041; Fax: (204) 474-7622
Canadian Association of Physicists (CAP)
Association canadienne des physiciens et physiciennes (ACP)
The Canadian Association of Physicists was founded in 1945 as a non-profit association
representing the interests of Canadian physicists. The CAP is a broadly-based national
network of physicists working in Canadian educational, industrial, and research settings.
We are a strong and effective advocacy group for support of, and excellence in, physics
research and education. We represent the voice of Canadian physicists to government,
granting agencies, and many international scientific societies. We are an enthusiastic
sponsor of events and activities promoting Canadian physics and physicists, including the
CAP's annual congress and national physics journal. We are proud to offer and continually enhance our web site as a key resource for individuals pursuing careers in physics
and physics education. Details of the many activities of the Association can be found at
http://www.cap.ca . Membership application forms are also available in the membership
section of that website.
L'Association canadienne des physiciens et physiciennes a été fondée en 1946 comme
une association à but non-lucratif représentant les intérêts des physicien(ne)s
canadien(ne)s. L’ACP est un vaste regroupement de physiciens oeuvrant dans les
milieux canadiens de l'éducation, de l'industrie et de la recherche. Nous constituons un
groupe de pression solide et efficace, ayant pour objectif le soutien de la recherche et de
l'éducation en physique, et leur excellence. Nous sommes le porte-parole des physiciens
canadiens face au gouvernement, aux organismes subventionnaires et à plusieurs
sociétés scientifiques internationales. Nous nous faisons le promoteur enthousiaste
d'événements et d'activités mettant à l'avant-scène la physique et les physiciens canadiens, en particulier le congrès annuel et la revue de l'Association. Nous sommes fiers d'offrir et de développer continuellement notre site Web pour en faire une ressource-clé pour
ceux qui poursuivent leur carrière en physique et dans l'enseignement de la physique.
Vous pouvez trouver les renseignements concernant les nombreuses activités de l’ACP à
http://www.cap.ca. Les formulaires d’adhésion sont aussi disponibles dans la rubrique
“Adhésion” sur ce site.
II
C PHYSICS
IN
CANADA / VOL. 64, NO. 4 ( Oct.-Dec. (Fall) 2008 )
Normand Mousseau
Chair du recherche du Canada, Département de physique
Université de Montréal, C.P. 6128, Succ. centre-ville
Montréal, Québec H3C 3J7
(514) 343-6614; Fax: (514) 343-2071
Michael Steinitz, PPhys
Department of Physics
St. Francis Xavier University, P.O. Box 5000
Antigonish, Nova Scotia B2G 2W5
(902) 867-3909; Fax: (902) 867-2414
Robert Thompson, PPhys
Dept. of Physics and Astronomy
University of Calgary, 2500 University Dr. NW
Calgary, Alberta T2N 1N4
(403) 220-5407; Fax: (403) 289-3331
ANNUAL SUBSCRIPTION / ABONNEMENT ANNUEL :
$40.00 Cdn + GST or HST (Cdn addresses),
$40.00 US (US addresses); $45.00 US (other/foreign addresses)
Advertising, Subscriptions, Change of Address/
Publicité, abonnement, changement d'adresse:
Canadian Association of Physicists /
Association canadienne des physiciens et physiciennes,
Suite/Bureau 112, Imm. McDonald Bldg., Univ. of/d' Ottawa,
150 Louis Pasteur, Ottawa, Ontario K1N 6N5
Phone/ Tél: (613) 562-5614; Fax/Téléc. : (613) 562-5615
e-mail/courriel : [email protected]; Website/Internet : www.cap.ca
Canadian Publication Product Sales Agreement No. 0484202/
Numéro de convention pour les envois de publications canadiennes :
0484202
© 2008 CAP/ACP
All rights reserved / Tous droits de reproduction réservés
WWW.CAP.CA
(select Physics in Canada /
Option : La Physique au Canada)
PRÉFACE
ASTROPHYSICS
U
nlike most other scientists, an astronomer does
are credited with discovering many moons of Jupiter,
not have direct access to the objects he is
Saturn, Uranus and Neptune, report on the often complex
researching. As a matter of fact, except for a few
interactions between Neptune and the thousands of objects
cases (solar wind and neutrinos, lunar samples,
that shape the Kuiper belt, of which Pluto is now only one
meteorites, …), all information originating from the
of the most massive components.
Universe is transmitted to us
Polarization is without conby light. Because it has the
test one of the less exploited
ability to interact with matter,
“ The marvel of marvels was that there on the rounded
properties of light in astronoit keeps a lasting impression
back of the planet, between this magnetic sheet and
my. With a relatively low
of the environment where it
those stars, a human consciousness was present in
level, therefore difficult to
was born or has had interacwhich as in a mirror that rain could be reflected. ”
measure, polarization detects
tion with. One of the greatest
the presence of a magnetic
challenges in astronomy is
Antoine de Saint-Exupéry, Wind, Sand and Stars
field in the interstellar medithus to extract, using methods
um, as demonstrated by
ever more clever, the maxiBrown and her colleagues in
mum information from photons that crossed through space
the article on one of the most important mappings of the
over thousands, if not billions, of years. A giant step forMilky Way, produced at the radio telescope in Penticton.
ward was taken nearly 400 years ago when Galileo Galilei
pointed a small and modest telescope towards the sky.
The next four articles touch on one of the hottest themes in
Technological developments have since considerably
contemporary astrophysics, the formation and evolution of
increased the dimension and visual acuity of telescopes
galaxies, but under completely different angles. First, Venn
(segmented mirrors, adaptive optics), the quantum efficienstates the importance of the spectroscopic study on metalcy of detectors (often close to 100%), the detectable wavepoor stars, witnesses to the first phases of development of
length range (from radio waves to gamma rays), as well as
the Milky Way and its neighbours. As Martel explains, the
all the specific measurement techniques such as photomeadvent of always more powerful computers and of innovatry, spectroscopy and polarimetry. Theoretical developtive algorithms allowing the simulation of the complex
ments are not overlooked with the always increasing use of
gravitational and hydrodynamic processes at play during
numerical modeling. Canada has been actively participating
collisions between galaxies help us better understand the
for a long time in the development of international astronomorphological, dynamical and chemical evolution of galaxmy, with its super computers, local and national infrastrucies. These numerical simulations are brought forth in the
tures (Mont Mégantic Observatory in the Eastern
results of a long term observation program taken on severTownships, DRAO in the Okanagan Valley, DAO in
al years ago by Abraham and his team through a new techVictoria, to name a few) or in collaboration with other
nique implemented at the Gemini telescope; this research
countries, on Earth (Canada-France-Hawaii, Gemini,
has made it possible for the first time to measure the propALMA telescopes) and in space (James Webb, MOST,
erties of galaxies when the Universe was only three to six
FUSE or UVIT telescopes).
billion years old. This article also demonstrates that the rate
of star formation in the Universe has dramatically dropped
To celebrate the upcoming International Year of Astronomy,
since that era and that this tendency will increase until the
this special issue offers an overview, very incomplete
depletion of all gas in the galaxies. Finally, the use of space
though, of Canadian astronomy and its research; some martelescopes to probe the heavy ultraviolet radiation emitted
ginal (but so fascinating!), other more conventional, but all
during star-forming bursts is well demonstrated by Robert,
leading edge in their respective field.
who reminds us of the role taken by the Canadian Space
Life on our planet depends on the sun’s luminosity, which
Agency in the development and commissioning of teleoriginates from nuclear reactions within its core. During the
scopes in space.
last 4 billion years, it has substantially increased, but also
From the solar system to distant galaxies, these articles
fluctuated to a lesser degree, depending on the intense magtouch on very diverse themes with theoretical and observanetic activity on the surface. The article by Charbonneau,
tional approaches, which we hope will fill the reader with
Crouch and Tapping demonstrates the modeling of these
wonder and food for thought.
solar irradiance variations.
At the root of the media frenzy which took place in 2006
following the withdrawal of Pluto as a planet by the
International Astronomical Union are the huge advances
realized during the last decade in recognizing the external
regions of the solar system. Gladman and Kavelaars, which
Laurent Drissen, Université Laval, Québec
Guest Editor, Physics in Canada
Comments of readers on this foreword are more than
welcome.
The contents of this journal, including the views expressed above, do not necessarily represent the views or policies of the
Canadian Association of Physicists. Le contenu de cette revue, ainsi que les opinions exprimées ci-dessus, ne représentent
pas nécessairement les opinions et les politiques de l’Association canadienne des physiciens et des physiciennes.
Laurent Drissen
<[email protected].
ca> est professeur au
département de
physique, de génie
physique et d'optique
de l'Université Laval. Il
est aussi titulaire de la
chaire de recherche
du Canada sur les
étoiles massives et
l'imagerie hyperspectrale depuis 2001.
Professor of Physics
at Laval University,
Laurent Drissen
<[email protected].
ca> has always been
fascinated by the most
massive stars. He
also works on the
development of imaging Fourier transform
spectrometers to
understand the properties of the ionized
interstellar medium.
LA PHYSIQUE AU CANADA / Vol. 64, No. 4 ( oct. à déc. (automne) 2008 ) C 199
FOREWORD
ASTROPHYSIQUE
C
ontrairement à la plupart des autres scientifiques,
Kuiper, dont Pluton n’est aujourd’hui qu’une des composantes
l’astronome n’a pas directement accès aux objets de
les plus massives.
ses recherches. En effet, à quelques exceptions près
(vent et neutrinos solaires, échantillons lunaires,
La polarisation est sans contredit l’une des propriétés de la
météorites, …), toute l’information en provenance de l’Univers
lumière les moins bien exploitées en astronomie. D’un niveau
nous est transmise par la lumière.
relativement peu élevé, et donc
Celle-ci, par sa capacité à interagir
difficilement mesurable, la polariavec la matière, garde une
sation permet cependant de
« Mais le plus merveilleux était qu’il y eut là, debout
empreinte indélébile du milieu qui
détecter la présence du champ
sur le dos rond de la planète, entre ce linge aimanté et
l’a vu naître ou avec lequel elle a ces étoiles, une conscience d’homme dans laquelle cette magnétique dans le milieu intereu une interaction. Un des plus
stellaire, comme le démontrent
pluie pût se réfléchir comme dans un miroir. »
grands défis de l’astronome conBrown et ses collègues dans leur
siste donc à extraire, par des
article de revue faisant état de
Antoine de Saint-Exupéry, Terre des Hommes
moyens plus ingénieux les uns que
l’une des plus importantes carles autres, le maximum d’informatographies de la Voie lactée, réaltion des photons ayant traversé l’espace pendant des milliers,
isée au radiotélescope de Penticton.
voire des milliards d’années. Un grand pas a été franchi il y a
près de 400 ans, alors que Galileo Galilei pointa une lunette, de
Les quatre autres articles abordent un thème parmi les plus
dimension modeste, vers le ciel. Depuis, le développement techchauds de l’astrophysique contemporaine, soit la formation et
nologique a permis d’augmenter considérablement la dimension
l’évolution des galaxies, mais sous des angles complètement
et l’acuité visuelle des télescopes (miroirs segmentés, optique
différents. Dans un premier temps, Venn relate l’importance de
adaptative), l’efficacité quantique des détecteurs (qui frôle soul’étude spectroscopique des étoiles pauvres en métaux, témoins
vent le 100%), le domaine observable de longueurs d’ondes (des
des premières phases de formation de la Voie lactée et de ses
ondes radio aux rayons gamma), sans oublier toutes les subtilités
voisines. L’avènement d’ordinateurs de plus en plus puissants et
des techniques de mesure que sont la photométrie, la spectrod’algorithmes innovateurs permettant de simuler les complexes
scopie et la polarimétrie. Le développement théorique n’est
processus gravitationnels et hydrodynamiques en jeu lors de colpas en reste, avec l’utilisation de plus en plus fréquente de la
lisions de galaxies nous permet de mieux comprendre l’évolumodélisation numérique. La Canada participe activement, et
tion morphologique, dynamique et chimique des galaxies,
depuis longtemps, au développement de l’astronomie internacomme nous l’explique Martel. Ces simulations numériques
tionale, que ce soit avec ses super-ordinateurs, ses infrastructrouvent écho dans les résultats d’un programme d’observation
tures locales et nationales (Observatoire du Mont Mégantic en
à long terme entrepris il y a plusieurs années par Abraham et son
Estrie, DRAO dans la vallée de l’Okanagan, DAO à Victoria,
équipe grâce à une nouvelle technique de spectroscopie mise en
pour ne nommer que ceux-là) ou en collaboration avec d’autres
place au télescope Gemini; cette recherche a permis pour la prepays, sur Terre (télescopes Canada-France-Hawaii, Gemini,
mière fois de mesurer les propriétés des galaxies à une époque
ALMA) et dans l’espace (télescopes James Webb, MOST, FUSE
où l’Univers n’avait que trois à six milliards d’années. Cet artiou UVIT).
cle nous démontre aussi que le taux de formation d’étoiles dans
l’Univers a dramatiquement chuté depuis cette époque et que
Afin de célébrer l’Année mondiale de l’astronomie qui est à nos
cette tendance ne fera que s’accentuer jusqu’à l’épuisement du
portes, nous vous proposons un numéro spécial consacré à un
gaz dans les galaxies. Finalement, l’utilisation de télescopes
tour d’horizon, très incomplet bien sûr, de l’astronomie canadispatiaux pour sonder le rayonnement ultraviolet copieusement
enne et ses axes de recherche; certains marginaux (mais combiémis lors des sursauts de formation stellaire est bien illustré par
en fascinants!), d’autres plus conventionnels, mais tous à la fine
Robert, qui nous rappelle aussi le rôle joué par l’Agence spatiale
pointe de leur domaine respectif.
canadienne dans le développement et la mise en service de télescopes dans l’espace.
La vie sur notre planète est tributaire de la luminosité du soleil,
qui tire son origine des réactions nucléaires en son coeur et qui
Du système solaire aux galaxies lointaines, ces articles abordent
a considérablement augmenté au cours des 4 derniers milliards
donc des thèmes très diversifiés avec des approches observad’années, mais qui fluctue aussi, à une moins grande échelle, au
tionnelles et théoriques; le lecteur y trouvera, nous l’espérons,
gré de l’intense activité magnétique qui règne en surface.
matière à réflexion et émerveillement.
L’article de Charbonneau, Crouch et Tapping fait état de la modélisation de ces variations de l’irradiance solaire.
Laurent Drissen, Université Laval, Québec
Rédacteur honoraire, La Physique au Canada
La tempête médiatique qui a fait rage en 2006 lorsque l’Union
astronomique internationale a retiré le statut de planète à Pluton
Les commentaires de nos lecteurs au sujet de cette préface sont
tire sa source des progrès immenses accomplis depuis une
bienvenus.
décennie dans notre connaissance des régions externes du système solaire. Gladman et Kavelaars, qui ont à leur crédit la
NOTE: Le genre masculin n’a été utilisé que pour alléger le
découverte de plusieurs lunes de Jupiter, Saturne, Uranus et
texte.
Neptune, nous font part des interactions souvent complexes
entre Neptune et les milliers d’objets qui forment la ceinture de
200 C PHYSICS
IN
ARTICLE DE FOND
L'IRRADIANCE SOLAIRE ET SES VARIATIONS
PAR
PAUL CHARBONNEAU, ASHLEY CROUCH,
ET
KEN TAPPING
L
’irradiance solaire totale est définie comme la
quantité d’énergie radiative originant du soleil,
intégrée sur toute les longueurs d’onde, incidente
sur un mètre carré de la haute atmosphère terrestre lorsque la terre est située à exactement une unité
astronomique (1.496 H 108 km) du soleil. Bien que les premières mesures de cette quantité fondamentale remontent
au dix-neuvième siècle, ce n’est que depuis une trentaine
d’années que les radiomètres et bolomètres spatiaux ont
atteint une stabilité suffisante pour en détecter les variations d’une manière fiable sur de longues échelles de
temps.
Fig. 1
La Figure 1 présente les variations observées de l’irradiance depuis 1978 (traits bleus). Cette séquence temporelle
est en fait un composite de mesures provenant de plusieurs
instruments différents [1]. On y voit l’irradiance varier sur
des échelles de temps allant de quelques heures jusqu’à la
décennie. Cette dernière variation est en phase avec le
cycle de l’activité magnétique du soleil, mieux connu en
termes de la variation quasi-cyclique du nombre de taches
solaires observées sur le soleil (voir Figures 2 et 3). Ceci
suggère que les fluctuations de l’irradiance sont associées
à celles du champ magnétique solaire, dont les taches
solaires sont un indicateur facilement observable.
Deux classes d’explications ont été avancées pour expliquer les fluctuations de l’irradiance. La première y voit
une simple conséquence de la variation, en fonction du
RÉSUMÉ
Nous décrivons ici une procédure de modélisation des variations de l'irradiance solaire
associées au cycle d'activité magnétique.
Notre approche, basée sur un modèle
physique simple plutôt que sur des corrélations statistiques établies sur des bases
purement observationnelles, permet en
principe une reconstruction physiquement
fiable des variations de l'irradiance durant
les siècles passés. Nous présentons une
reconstruction remontant jusqu'à 1874,
incluant une modulation à long terme de l'irradiance de la partie non-magnétisée de la
photosphère solaire, cette modulation étant
calculée à l'aide d'un modèle semi-empirique
de la variation du flux magnétique total à l'intérieur du soleil établi à l'aide des archives
du flux solaire radio F10.7.
Variations observées de l’irradiance solaire totale
depuis 1978. Le trait bleu clair correspond aux valeurs
journalières, et le trait bleu foncé une moyenne courante
de largeur 81 jours. Ces données sont tirées du composite d41_61_0702 distribué par le PhysikalischMeteorologisches Observatorium Davos. Les traits
orange et rouge sont leurs équivalents tels que produit
par les simulations décrites plus bas, artificiellement
décalés vers le bas de 4 W/m2 afin de faciliter la comparaison.
cycle d’activité, de la fraction du disque solaire occuppée
par des structures magnétiques ayant des émissivités
radiatives différentes des régions non-magnétisées de la
photosphère [2-4]. On sait, par exemple, que les structures
magnétiques de plus grandes tailles, comme les taches
solaires, nuisent au transport convectif de l’énergie, et se
retrouvent donc plus froides que la photosphère, ce qui
cause un déficit d’irradiance. Par contre, les soi-disantes
facules et éléments du réseau supergranulaire, structures
magnétisées de bien plus petites tailles mais beaucoup
plus nombreuses, produisent un déficit local de densité au
niveau de la photosphère, permettant ainsi de “voir” les
régions plus profondes — et donc plus chaudes — du
soleil. Collectivement, l’ensemble de ces petites structures
magnétiques conduit donc à un excès d’irradiance qui
dépasse légèrement le déficit associé aux taches, ce qui se
traduit en un soleil légèrement plus brillant au maximum
d’activité (cf. Figures 1 et 3). Il est maintenant bien
démontré que les variations de l’irradiance sur les courtes
échelles temporelles (de la minute à l’année) sont bel et
bien associées à ces effets de surfaces [5]. Une seconde
classe d’explication attribue une partie, voire la totalité,
des variations sur les échelles décadales et plus à une modulation de l’efficacité du transport convectif de l’énergie
depuis la base de la zone convective solaire, 200000 km
sous la surface, causée par une interaction dynamique
entre le fort champ magnétique agissant comme moteur du
cycle d’activité, et les écoulements fluides associés à la
Paul Charbonneau
<paul.charbonneau@
umontreal.ca> et
Ashley Crouch,
Département de
physique, Université
de Montréal, C.P.
6128 Succ. CentreVille, Montréal, QC,
H3C 3J7, CANADA;
et
Ken Tapping,
Institut Herzberg
d'Astrophysique,
Conseil national de la
recherche, P.O. Box
248, Penticton, C-B
V2A 6J9, CANADA
L’IRRADIANCE SOLAIRE ... (CHARBONNEAU ET AL.)
niveau des taches solaires (voir
Figure 3). Comment se comporte
l’irradiance durant des phases
d’augmentation graduelle du
niveau général d’activité, comme
durant la première moitié du
vingtième siècle, ou durant des
périodes prolongées d’activité
fortement réduite, comme le
fameux Minimum de Maunder
entre 1645 et 1715?
convection [6,7]. Il va sans
dire que ces deux classes
d’explications ne sont pas
mutuellement exclusives.
Il n’en demeure pas moins
que depuis 1978, les variations
de
l’irradiance
observées entre les phases
maximales et minimales du
cycle d’activité sont très
faibles, de l’ordre de 0.1%.
Une si faible variation, sur
une échelle décadale de
surcroit, n’a essentiellement pas d’effet sur le climat terrestre, en raison
principalement de l’inertie
thermique que confèrent les
océans au climat. On doit
cependant remarquer que
les derniers trois cycles
d’activité, les seuls pour
lesquels nous disposons de
mesures directes et fiables
de l’irradiance, ne sont pas
particulièrement représentatifs des variations de l’activité solaire observées
depuis le début du dix-septième siècle, du moins au
Fig. 3
Fig. 2
Image du soleil en lumière continue, où l’on peut noter la
présence de nombreuses taches solaires, plus sombres que la
photosphère, et de facules, structures filamentaires plus brillantes que la photosphère et surtout visible près du limbe, où
leur contraste est plus grand. Les éléments brillants du réseau
supergranulaire sont trop petits pour être visibles à cette résolution spatiale. Image obtenue le 30 mars 2001 par l’imageur
optique sur SOHO/MDI (NASA), en phase élevée du cycle
d’activité magnétique.
La coincidence temporelle entre le
Minimum de Maunder et la phase la
plus marquée du “petit âge
glaciaire” bien connu en climatologie (voir Ref. [9], et références s’y
trouvant) a motivé une vaste
gamme de tentatives d’estimation
du niveau auquel aurait pu chuter
l’irradiance solaire durant une telle
phase prolongée d’inactivité, et,
plus généralement, de l’impact de
l’activité solaire sur le climat terrestre. Ces questions sont d’autant
plus importantes que les études du
niveau général de l’activité solaire
via la mesure des abondances de
radioisotopes cosmogéniques [10]
ont démontré hors de tout doute
que le Minimum de Maunder n’est
Variation du nombre de taches solaires observées à la surface du soleil en fonction du temps. Il s’agit ici du Nombre de Wolf mensuel
(orange) et annuel (rouge), tel que distribué par le Solar Influences Data Analysis Center (http://sidc.oma.be). Les cycles sont numérotés
d’après la convention introduite au 19ème siècle par Rudolf Wolf. Le trait vert correspond au “Group SunSpot Number” de Hoyt & Schatten
[8], généralement considéré plus fiable que le nombre de Wolf avant 1750. Quelle que soit la mesure utilisée, on note un cycle bien défini, d’amplitude variable et d’une période allant de 9 à 14 ans, avec une valeur moyenne de 11 ans. Le cycle magnétique sous-jacent a une
période du double du cycle des taches, ces dernières se formant quelle que soit la polarité du champ magnétique solaire interne. Le peu de
taches solaires observées durant le Minimum de Maunder (1645-1715), ne reflète pas un manque de données mais représente un épisode
d’activité fortement réduite. Les traits verticaux verts délimitent la période 1978-2007, pour laquelle des mesures de l’irradiance sont
disponibles (Figure 1).
202 C PHYSICS
IN
L’IRRADIANCE SOLAIRE ... (CHARBONNEAU ET AL.) AA
pas un événement unique, mais que de tels épisodes se sont
produits de manière irrégulière une trentaine de fois dans
les derniers 10000 ans [11], et donc pourrait fort bien se produire de nouveau.
La majorité des reconstructions de l’irradiance remontant
jusqu’au dix-septième siècle publiées à date sont basées sur des
corrélations statistiques établies à l’aide d’observations magnétographiques récentes, permettant d’établir un lien entre la couverture surfacique de diverses classes de structures photosphériques magnétisées, et des indicateurs indirects ayant une
longue archive temporelle comme les taches solaires, le
niveau d’émission dans certaines raies spectrales sensibles à
la présence d’un champ magnétique, le flux radio F10.7,
etc. [12-15]. Une approche alternative consiste à établir un lien
évolutif entre les taches, facules, réseau, etc., via un modèle
physique simple. La reconstruction de l’irradiance en résultant
doit aussi être calibrée sur 1978-2007, mais l’universalité des
lois de la physique suggère qu’un tel modèle peut être
extrapolé avec plus de confiance qu’une combinaison de corrélations de nature purement statistique. L’incertitude provient
maintenant principalement du niveau de réalisme du modèle
physique utilisé.
Dans ce qui suit, nous décrivons un modèle physique de ce
genre, récemment développé par le groupe de recherche en
physique solaire à l’Université de Montréal [16,17] (voir également Refs. [18] et [19] pour des reconstructions basées en
partie sur la même “philosophie”). Nous discutons ensuite
l’étalonnage de ses paramètres internes à partir des données de
l’irradiance pour 1978-2007, et présentons finalement une
reconstruction de l’irradiance remontant à 1874 produites par
le modèle ainsi calibré, incluant une variation à long terme du
niveau général d’activité.
MODÉLISATION PHYSIQUE DE L’IRRADIANCE
Notre procédure de modélisation de l’irradiance est décrite
avec abondance de détails dans Ref. [16], ce qui suit n’étant
qu’un bref résumé. Le simulation couvre la surface du soleil, et
évolue dans le temps selon un pas journalier. Jour après jour,
les taches solaires sont “injectées” dans la simulation aux
latitudes, longitudes et avec la taille correspondant aux observations d’émergence de taches et régions actives compilées à
partir des données photographiques du Royal Greenwich
Observatory pour 1874-1976, et de l’USAF pour 1977-2007
(voir
http://solarscience.msfc.nasa.gov/greenwch.shtml).
Comme nous ne pouvons observer que la moitié de la surface
du soleil, on s’attend à ce que cette banque de données ne contienne que la moitié des émergences s’étant produites à la surface du soleil; nous introduisons donc une procédure statistique
stochastique afin de modéliser les émergences sur la face
cachée.
Les taches ainsi injectées se désagrègent par la suite en plus
petites structures, sous l’effet de la fragmentation et de l’érosion en leur périphérie (processus pour lesquels il existe un bon
support observationnel; voir, e.g., Refs. [20-22], et références
s’y trouvant). Ces fragments deviennent par la suite sujets au
même processus de fragmentation/érosion, jusqu’à une certaine taille minimale en deça de laquelle les fragments disparaissent (e.g. par submergence convective) avec une probabilité dont la valeur, difficile à contraindre observationnellement, est traitée comme un paramètre du modèle.
La procédure décrite ci-dessus produit une distribution de
“fragments” de tailles diverses, évoluant dans le temps. Ces
fragments sont ensuite regroupés en deux grandes classes selon
leurs tailles, soit les “taches”, observées comme étant plus
sombres que la photosphère, et les “facules”, plus brillantes.
Dénotant par SQ l’irradiance de la photosphère non-magnétisée
(incluant l’effet du noircissement centre-bord), on modélise
l’irradiance selon l’expression
S (t ) = SQ +
N s (t )
N f (t )
i =1
j =1
∑ ΔSs,i +
∑ ΔS
f,j
,
(1)
où Ns (t) et Nf (t) correspondent au nombre de taches et facules
présentes en surface au jour t, ΔSs,i correspond au déficit d’irradiance associé à la tache i, et ΔSf, j à l’excès correspondant à
la facule j, Ces deux dernières quantités dépendent à la fois de
la taille et de la position de la structure à la surface du soleil;
nous utilisons les expressions suivantes, établies empiriquement (voir, e.g., [Refs. 4, 23, et 24]):
ΔS s ,i
SQ
1
= − μ As ,i (3μ + 2)
2
ΔS f , j
SQ
(2)
⎛1 ⎞
1
= − μ Af , j (3μ + 2) ⎜ − 1⎟ α f ,
2
⎝μ ⎠
(3)
où les couvertures surfaciques As,i et Af,j sont exprimées en
fraction de l’hémisphère (2πR 2), μ = cos θ cos φ mesure l’angle centre-bord, et le contraste faculaire αf est considéré
comme paramètre du modèle, ses déterminations observationnelles n’étant pas très précises. Notons que la couverture surfacique totale des taches est donnée par
As (t ) =
N s (t )
∑A
s ,i
.
(4)
i =1
Il est crucial de noter que ce modèle, si simple soit-il, est de
nature partiellement stochastique: la même séquence de données en entrée (émergences de taches) produira des courbes
d’irradiance distinctes sous diverses réalisations aléatoires du
processus de fragmentation et/ou des émergences sur la face
cachée du soleil. Il s’avère que ce second aspect domine la stochasticité du modèle (pour plus de détails voir Ref. [16]).
ÉTALONNAGE DU MODÈLE: 1978-2007
En bout de ligne, le modèle de base décrit ci-dessus se retrouve défini par 10 paramètres, dont seulement quatre peuvent être
contraints de manière fiable sur la base des observations. Les
six autres paramètres doivent donc être ajustés de manière à
offrir la meilleure représentation possible de l’irradiance
observée dans l’intervalle 1978-2007. À prime abord il s’agit
ici d’un problème d’optimisation classique, mais la stochasticité inhérente au modèle complique sérieusement la situation,
tout comme le fait que l’espace des paramètres est fortement
multimodal, i.e., plusieurs combinaisons distinctes de
paramètres peuvent produire des séquences temporelles d’irradiance comparables.
Face à ces difficultés, nous avons choisi d’optimiser le modèle
à l’aide d’un algorithme génétique (ci-après AG), plus spécifiquement la version 1.2 du logiciel PIKAIA [25,26]. Nous cherchons à minimiser simultanément les écarts quadratiques
moyens entre les séquences temporelles S(t) et As(t) observées,
et celles produites par le modèle décrit ci-dessus. Voir la §3 de
Ref. [16] pour tous les détails concernant cette procédure de
minimisation.
Une solution typique S(t) est présentée sur la Figure 1, sous une
forme équivalente aux observations (trait orange: valeurs journalières; trait rouge: moyenne courante sur 81 jours). En raison
des aspects stochastiques du modèle, on ne peut s’attendre à ce
que les observations soient reproduites dans tous leurs détails;
cependant il est clair que la simulation reproduit bien l’allure
générale des variations observées, autant au niveau des amplitudes que des échelles de temps. Considérant la simplicité (relative) du modèle dans sa forme actuelle, le résultat est remarquable!
imisations par AG, débutant de populations initiales distinctes
et utilisant des réalisations aléatoires distinctes des émergences
sur la face cachée et de la fragmentation des taches. Travaillant
encore une fois à partir des séquences temporelles lissées sur
81 jours, pour chaque pas de temps (journalier) on calcule la
moyenne (⎯S) des 1000 simulations ainsi que la déviation standard (σ) par rapport à cette moyenne; on trace ensuite un trait
vertical (rouge) couvrant l’intervalle⎯S ± σ. La répétition de
cette procédure à chaque pas de temps journalier produit la
bande rouge. La même procédure, appliquée cette fois à la couverture surfacique totale des taches, As(t), est illustrée en (b).
La courbe verte en (a) montre la variation sinusoidale correspondant à une des meilleures solutions produites par l’AG.
Ces simulations présentent des résidus quadratiques moyens de
l’irradiance significativement plus petits (0.168 W m-2 plutôt
que 0.202 W m-2) que pour un niveau basal SQ constant,
comme dans l’éq. (1). Ceci suggère qu’il existe une source de
structures “brillantes” qui n’est pas reliée à la désagrégation
des taches solaires, mais qui varie néanmoins approximativement en phase avec le cycle d’activité. Les observations
solaires offrent déjà plusieurs pistes quant à la nature de cette
source; on observe ainsi souvent des facules émergeant simul-
La flexibilité de l’AG nous permet de généraliser facilement le
modèle de manière à inclure des sources additionelles d’irradiance, n’étant pas directement associées à la désagrégation des
taches. Par exemple, certaines observations suggèrent qu’il
existe une source de structures de type faculaire qui contribuent
un excès d’irradiance en début de cycle. Les observations historiques de l’activité solaire (Fig. 3) indiquent aussi la présence
de variations à long terme du niveau général de l’activité, qui
pourraient s’accompagner de variations du niveau de base de
l’irradiance.
Considérons tout d’abord l’existence possible d’une variation
cyclique de l’irradiance du disque (SQ dans l’éq. (1)), i.e., le
modèle d’irradiance est maintenant défini par:
S (t ) = SQ + S0 sin ( ωt + φ ) +
N s (t )
∑ ΔSs,i +
i =1
N f (t )
∑ ΔS
f,j
.
(5)
j =1
La fréquence angulaire ω, la phase φ et l’amplitude S0 de cette
variation sont toutes traitées comme des paramètres libres,
ajustés via l’AG simultanément aux autres paramètres du modèle de base décrit à la §2 (voir Ref. [17] pour plus de détails
sur ces modèles “améliorés”).
Fig. 4
La Figure 4 présente les résultats produits par un tel modèle
une fois calibré à l’intervalle 1978-2007, dans un format permettant de quantifier l’impact des aspects stochastiques du
modèle. La partie (a) présente la courbe d’irradiance journalière observée (trait bleu), lissée via une moyenne courante de
81 jours. La bande rouge représente les résultats de 1000 min-
204 C PHYSICS
IN
Séquences temporelles des moyennes courantes (largeur 81
jours) de (a) l’irradiance, et (b) la couverture surfacique totale
des taches. Les traits bleus correspondent aux observations
(cf. Fig. 1), et les bandes rouges à la moyenne ± une déviation standard de 1000 reconstructions utilisant des réalisations distinctes des émergences sur la face cachée. Le trait
vert en (a) correspond à la variation sinusoidale du niveau de
base de l’irradiance (partie non-magnétisée de la photosphère), incluse dans ces simulations selon l’éq. (5). Pour
cette solution, la variation sinusoidale contribue presque
autant que les facules à l’excès d’irradiance compensant le
déficit associé aux taches.
tanément (ou parfois même un peu avant) les taches produisant
une nouvelle région d’activité. Une explication alternative
assignerait la variation sinusoidale à une variation structurelle
de la zone convective solaire. Il n’est présentement pas possible de démarquer ces différents scenarii uniquement sur la base
de nos résultats de modélisation. Cependant, les mesures
simultanées de l’irradiance, du diamètre solaire et de la forme
du limbe solaire qui seront produites par la mission spatiale
PICARD (http://earth-sciences.cnes.fr/PICARD/Fr/), dont le
lancement est prévu pour juin 2009, pourraient fort bien changer la donne.
RECONSTRUCTION DE L’IRRADIANCE
DEPUIS 1874
Une fois les paramètres libres du modèle fixés par étalonnage
sur l’intervalle 1978-2007, il devient possible d’utiliser le modèle pour produire des reconstructions de l’irradiance débutant
là où commence notre banque de données des émergences de
taches solaires, soit 1874. L’intervalle de temps ainsi simulé
couvre la majorité de l’ère industrielle, et en particulier la période de réchauffement terrestre global ayant débuté au
vingtième siècle. Le niveau général de l’activité a considérablement varié depuis 1874 (cf. Fig. 3), et ceci pourrait fort
bien produire une variation à long terme de SQ se superposant
aux variations associées à la couverture surfacique des taches,
facules, etc. La Figure 5 présente une reconstruction remontant
à 1874, incorporant une modulation à long terme de la contribution à l’irradiance provenant de la partie non-magnétisée de
la photosphère, calculée selon la procédure semi-empirique
décrite dans Tapping et al. [19]. Plus spécifiquement, l’éq. (1)
est maintenant remplacé par
viron 0.4 W m-2 entre 1900 et 1986; cette hausse est petite,
mais sur de telles échelles de temps aurait déjà une influence
détectable sur le climat.
Il est clair que le détail du modèle d’irradiance et de la procédure d’étalonnage sur 1978-2007 peuvent avoir un impact
important sur les reconstructions de l’irradiance. Notons, par
exemple, que Tapping et al. [19], à partir du même profil de
qu’adopté ici mais utilisant une procédure de modélisation différente, arrive à un écart 1874-2008 de 0.8 W/m2, soit le double de celui caractérisant la reconstruction de la Figure 5. De
plus, d’autres reconstructions récentes (e.g. Ref. [28]) suggèrent une augmentation de l’irradiance depuis 1874 dépassant
2 W m-2. Il est donc impératif de développer des modèles des
variations 1978-2007 qui soient les plus réalistes possibles au
niveau de la physique sous-jacente, mais tout en demeurant
assez simples pour pouvoir servir de base à des simulations
couvrant plusieurs siècles, et pouvant être effectuées en un
temps de calcul raisonnable.
DU PAIN SUR LA PLANCHE...
Nous travaillons présentement à plusieurs améliorations de la
procédure de modélisation de l’irradiance solaire décrite cidessus, notamment au niveau de l’inclusion d’un modèle de
type agrégation/diffusion pour la formation des facules et des
éléments du réseau magnétique supergranulaire [29]. Nous
avons déjà entâmé la généralisation de la procédure à la modélisation de l’irradiance spectrale, soit la synthèse du spectre
solaire en entier, et plus particulièrement la portion du spectre
ultraviolet contrôlant la chimie et la dynamique de la
S (t ) = SQ ,c + SQ ,a S10.7 (t ) +
N s (t )
∑ ΔS
i =1
s ,i
+
N f (t )
∑ ΔS
f,j
.
(6)
j =1
où S10.7 est la variation lente du flux radio F10.7 en fonction du
temps, telle que reconstruite par Tapping et al. [19]. Le flux
radio F10.7 est généralement considéré comme étant un bon
indicateur du niveau général de l’activité magnétique globale [27], et parmi tous les indicateurs de l’activité solaire est
celui qui corrèlle le mieux avec l’irradiance. Le nouveau
paramètre SQ,a et le niveau basal constant SQ,c sont tous deux
ajustés via l’AG simultanément aux autres paramètres du modèle d’irradiance. Encore une fois, les meilleures solutions
évoluées par l’AG utilisant l’éq. (6) présentent de plus faibles
résidus quadratiques moyen (0.188 W m-2) que ceux caractérisant le modèle de base décrit par l’éq. (1).
Une reconstruction semblable mais utilisant le modèle de base
décrit à la section sur la modélisation physique de l’irradiance
(éq. (1) est présentée dans Ref. [16]; voir Figure 13). Le niveau
de l’irradiance aux minima d’activité d’une telle reconstruction
demeure fixé, par construction, à la valeur produite par la
procédure d’étalonnage du modèle sur 1978-2007, soit 1365.42
W/m2. La reconstruction de la Figure 5 ci-dessus, cependant,
accuse une hausse de l’irradiance aux minima d’activité d’en-
Fig. 5
Reconstruction des variations de l’irradiance solaire depuis
1874, calculée à l’aide de notre modèle. Cette reconstruction
incorpore une modulation à long terme reliée au flux magnétique total, pour lequel le flux radio F10.7 sert d’indicateur,
via l’éq. (6). Comme sur la Fig. 4, le trait rouge correspond
encore une fois à la moyenne de 1000 réalisation du modèle
optimal ± une déviation standard, et le trait orange fin à une
solution représentative extraite de cet ensemble. Le trait
bleu correspond au niveau basal d’irradiance (SQ = 1365.42
W m-2 dans l’éq. (1)) produit par la version du modèle n’incluant pas de modulation à long terme (cf. Fig. 1, traits
orange/rouge) et produisant une solution optimale sur l’intervalle 1978-2007. La hausse rapide de l’irradiance entre 1874
et 1875 est un transient artificiel associé à la condition initiale
utilisée, soit un soleil sans taches.
stratosphère terrestre. Nous continuons d’explorer l’utilisation
des archives du flux radio F10.7, remontant à 1947, comme
tremplin permettant une modélisation physique plus réaliste
des variations à long terme de la contribution à l’irradiance
provenant des régions non-magnétisées de la photosphère
solaire, ce qui nous permettrait de pousser les reconstructions
au moins jusqu’au Minimum de Maunder. À ce niveau il est
également intéressant de noter qu’au moment d’écrire ces
lignes (juin 2008), en phase minimale d’activité séparant le
cycle 23 du cycle 24, le flux radio F10.7 ainsi que l’irradiance
ont atteint des valeurs plus basses que jamais observées auparavant, et le début du cycle 24 accuse près d’un an de retard par
rapport aux prévisions faites durant la phase descendante du
cycle 23. Le soleil nous réserve-t-il une surprise? A suivre!
REMERCIEMENTS
Les travaux décrits ici bénéficient du support financier du
Conseil de Recherche en Sciences Naturelles et en Génie
(Programmes “Subventions à la Découverte” et “Chaires de
Recherche du Canada”), de la Fondation Canadienne pour
l’Innovation, et des Fonds Québécois de la Recherche sur la
Nature et les Technologies (Programme “Projet de Recherche
en Équipe”).
BIBLIOGRAPHIE
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11
12.
13.
14.
15.
16.
17.
18.
19.
20.
21.
22.
23.
24.
25.
26.
Fröhlich, C., & Lean, J., Astron. Astrophys. Rev., , 273-320, (2004).
Foukal, P., & Lean, J., Astrophys. J., 302, 826-835, (1986).
Chapman, G.A., Cookson, A.M., & Dobias, J.J., J. Geophys. Res., 101, 13541-13548, (1996).
Lean, J.L., Cook, J., Marquette, W., & Johannesson, A., Astrophys. J., 492, 390-401, (1998).
Foukal, P., & Bernasconi, P.N., Sol. Phys., 248, 1-15, (2008).
Kuhn, J.R., & Stein, R.F., Astrophys. J., 463, L117-L120, (1996).
Li, L.H., Basu, S., Sofia, S., Robinson, F.J., Demarque, P., & Guenther, D., Astrophys. J., 591, 1267-1284, (2003).
Hoyt, D.V., & Schatten, K., Sol. Phys., 181, 491-512, (1998).
Eddy, J.A., Science, 192, 1189-1202, (1976).
Beer, J., Space Sci. Rev., 94, 53-66, (2000).
Usoskin, I.G., Solanki, S., & Kovaltsov, G.A., Astron. Astrophys., 471, 301-309, (2007).
Foukal, P., & Lean, J., Science, 247, 505-604, (1990).
Hoyt, D.V., & Schatten, K., J. Geophys. Res., 98, 18895-18906, (1993).
Lean, J.L., Beer, J., & Bradley, R., Geophys. Res. Lett., 22, 3195-3198, (1995).
Krivova, N.A., Balmaceda, L., & Solanki, S.K., Astron. Astrophys., 467, 335-346, (2007).
Crouch, A.D., Charbonneau, P., Beaubien, G., & Paquin-Ricard, D., Astrophys. J., 677, 723-741, (2008).
Crouch, A.D., & Charbonneau, P., Astrophys. J., soumis, (2008).
Wang, Y.-M., Lean, J.L., & Sheeley, N.R.Jr., Astrophys. J., 625, 522-538, (2005).
Tapping, K.F., Boteler, D., Charbonneau, P., Crouch, A.D., Manson, A., & Paquette, H., Sol. Phys., 246, 309-326, (2007).
Schrijver, C.J., Astrophys. J., 547, 475-490, (2001).
Petrovay, K., & Moreno-Insertis, F., Astrophys. J., 485, 398-408, (1997).
Martinez Pillet, V., Astron. Nach., 323, 342-348, (2002).
Chapman, G.A., & Meyer, A.S., Sol. Phys., 103, 21-31, (1986).
Brandt, P.N., Stix, M., & Weinhardt, H., Sol. Phys., 152, 119-124, (1994).
Charbonneau, P., Astrophys. J. Suppl., 101, 309-334, (1995).
Charbonneau, P., Release Notes for PIKAIA 1.2, NCAR Technical Note TN-451-STR, Boulder: National Center for Atmospheric
Research, (2002).
27. Tapping, K.F., J. Geophys. Res., 92, 829-838, (1987).
28. Solanki, S.K., & Fligge, M., Geophys. Res. Lett., 25, 341-344, (1998).
29. Crouch, A.D., Charbonneau, P., & Thibault, K., Astrophys. J., 662, 715-729, (2007).
206 C PHYSICS
IN
ARTICLE DE FOND
RESONANCE DYNAMICS
BY
IN THE
KUIPER BELT
BRETT GLADMAN AND J.J. KAVELAARS
C
elestial mechanics is the queen of physics, being
the longest-studied quantitative subject. The
desire to predict the orbital motion of the planets
drove both mathematics and physics to develop
varied and precise techniques. For example, Newton’s
crowning achievment was the derivation of Kepler’s laws
of orbital motion from the three laws of motion and the
law of universal gravitation. Bessel derived his famous
functions to attempt to solve Kepler’s nonlinear equation
for the angular position of a planet along its orbit.
Kepler discovered (and stated in his First Law) that the
planetary orbits are well described by ellipses in a fixed
plane around the Sun, and that around the eccentric orbit
the speed varies. Beginning students all learn Kepler’s
Third law in its simple form
3
Pyr2 = a AU
(1)
for objects in orbit around the Sun, where the unit of distance is scaled to the Earth’s orbital semimajor axis of
1 AU 1.5 H 108 km. The semimajor axis a is one of 6
parameters (called ‘orbital elements’), which describe the
shape of the orbit (through a and the eccentricity e), its orientation (the inclination i, the longitude of ascending node
Ω, and the position of the perihelion ω relative to the
ascending node), and finally the angular position f along
the orbit from the perihelion point. In the perturbation
equations of celestial mechanics, the orbits of each of the
planets slowly change due to the weak tugs from the other
planets. In the classical secular perturbation theory
approach to compute the evolution of the orbital elements,
orbital resonances have an important role. The location of
a so-called mean-motion resonance is easy to compute; if
a planet has period P1, then an external resonance with an
second object whose orbital period P2 > P1 will occur
when P2/P1 is a ratio of two integers m/n, whereupon
Eq. (1) provides the resonant semimajor axis. For example, the external 5:2 mean-motion resonance of Neptune
30 AU) occurs where a2/a1 = (P2 / P1)2/3 =
(with a1
2/3
(5/2) , corresponding to a2 55 AU. A transneptunian
object (abbreviated TNO) with 55 AU will circle the Sun
twice for every 5 orbits of Neptune. In analogy with the
way amplitude can be built up by correctly timing the
pushes of a child on a swing, the repetitive geometrical
configuration caused by the resonant configuration results
in the resonance affecting the orbital evolution of the TNO
more than other (non-resonant) semimajor axes nearby.
The importance of mean-motion resonances in the Kuiper
Belt is easily seen in the semimajor axis/eccentricity distribution shown in Figure 1. The most prominent group
inhabits the 3:2 mean-motion resonance with a 39.4 AU;
these objects are known as ‘plutinos’ because Pluto was
the first object known to occupy this resonance. One can
see that most of the resonant objects have higher orbital
eccentricities than the main part of the Kuiper Belt (the
‘classical’ objects which range mostly from a = 38 – 48
AU); there are complex observational selection effects at
play here, and this figure must be interpreted with care.
What is clear is that resonant objects exist, and one naturally proposes the question as to how these objects entered
SUMMARY
Over the last decade it has become clear that
the Solar System’s Kuiper Belt has a rich
dynamical structure. Mean-motion resonances (which are depleted in the asteroid
belt and appear as the Kirkwood gaps in the
asteroidal semimajor axis distribution) are
preferentially populated in the Kuiper Belt,
confering orbital stability to objects which
would otherwise have short lifetimes against
gravitational encounters with Neptune. The
basics of the orbital mechanics of these resonances is presented. How planet migration
may be at the origin of the observed resonant structure is outlined, along with some
of the observational complications involved.
Fig. 1
The inner portions of the trans-neptunian region in
semimajor axis/eccentricity space, for TNOs with highquality orbits. Vertical lines show mean-motion resonance locations, along which many of the known TNOs
sit; open squares denote TNOs whose resonant arguments are known to oscillate. This figure is an update of
an orbital classification presented in Ref. [13].
B. Gladman
<[email protected].
ca>, Department of
Physics and Astronomy, Institute for
Planetary Science,
University of British
Columbia, 6224
Agricultural Road,
Vancouver, BC V6T
1Z1 Canada
and
J.J. Kavelaars
<[email protected].
ca>, Herzberg Institute
of Astrophysics,
National Research
Council of Canada,
5071 West Saanich
Road, Victoria, BC
V9E 2E7 Canada
RESONANCE DYNAMICS ... (GLADMAN/KAVELAARS)
this dynamical state. We must first understand a little bit better
what the resonant state actually is.
THE PENDULUM ANALOGY.
The language of orbital resonances share some terminology
with the more familiar problem of the rigid-rod pendulum,
where a mass m pivots on a rod of length L in a vertical plane,
with φ measuring the angular displacement from “straight
down”. The dimensional Hamiltonian of the problem is written
H dim =
p2
+ mgL(1 − cos φ)
2m
(2)
where g is the local acceleration with respect to gravity. The
single coordinate φ and its corresponding (generalized)
momentum p describe the motion. One can choose units of
mass, length, and time so that m = g = L = 1, giving the nondimensional form:
p2
+ (1 − cos φ)
2
with corresponding Hamiltonian equations of motion
∂H nd
φ=
= p,
∂p
H nd =
p=−
∂H nd
= − sin φ ,
∂φ
(3)
(4)
(5)
where the over-dot indicates differentiation with respect to
time. The reader will note that differentiating the first equation
and substituting for ṗ from the second
yields the second-order differential equation φ̈ + sinφ = 0 which, for small-amplitude oscillations which keep sin φ . φ,
produces the familiar simple harmonic
oscillator for small amplitude oscillations
around the straight-down point.
the evolution, a familiar fact that is verifiable by computing Ḣ.
Therefore, a curve of constant H is a trajectory. The figure
shows the small-amplitude trajectories surrounding the vertically-down (θ = π) point as circles, which become elongated as
motions that approach θ = 0 or π (the ‘straight-up’ point). The
thick curve connecting (θ, p) = (0,0) with (π,0) is called the
separatrix because in separates the oscillatory regime, centered
on (π,0), from the rotor regime at higher momenta (where the
pendulum circulates ‘over the top’ each period).
The terminology as it relates to celestial mechanics is closer if
one now moves to a new ‘action’ I which is just a linear offset
of p (a new inertial reference frame) and plots the trajectories
in polar coordinates (Fig. 3). Here the ‘resonant’ regime (of
oscillatory motion) is positioned at left within the thick separatrix and surrounding the low amplitude oscillations. The ‘resonant angle’ θ is the polar position of a vector starting from
the origin that points to the trajectory in question (determined
by value of H determined by the initial values of I and θ).
This vector moves along a surface of constant H, controlling
the value of the polar angle. For oscillatory motions inside the
separatrix, θ does not explore all values, but has a restricted
range centered on
180o. Recall that
we have not determined the time
behaviour along
the
trajectories
(which is much
For what follows it is simpler to use an
angular coordinate θ = π - φ which is the
angular position in the plane of oscilation, but measured from the ‘straight-up’
point, so that θ = π is the equilibrium
point at the bottom of the swing. The
Hamiltonian is
H=
p2
+ cos θ
2
(6)
The full solutions for θ(t) and p(t) of the
simple-looking equations of motion analogous to (4) and (5) require knowledge of
the Jacobian elliptic functions. However,
if one only wishes to understand the qualitative nature of the trajectories one can
simply plot level sets of the Hamiltonian
H in the (θ,p) plane (Fig. 2). This is
because the value of H (the system’s
mechanical energy) is conserved during
208 C PHYSICS
IN
Fig. 3
Fig. 2
The phase space of a planar rigid-rod
pendulum, viewed in cylindrical coordinates. The angle θ measures the angle
down from the vertically straight-up
position, while p is the (angular)
momentum in the intertial frame where
the pivot is at rest. Trajectories (lines of
constant H) are shown. The oscillatory
(also called librating) trajectories are
dashed, while the ‘rotor’ solutions
(which visit all values of θ) are solid.
The thick trajectory is the separatrix.
The dynamics viewed in polar coordinates. The state of the system is
expressed by the polar angle θ
(measured counterclockwise from
the dashed reference direction) and a
generalized momentum I which
determines the length of the state
vector. Curves are still the level sets
of the Hamiltonian and the thick
curve is the separatrix of the previous figure. Rotor motions (solid trajectories) are both exterior to the separatrix and also near the origin (corresponding to ‘above’ and ‘below’
the separatrix in the previous figure).
Librating trajectories are again
dashed. As the system evolves along
one of the trajectories the angle θ
changes. In the case shown θ will not
take on all values but rather will
oscillate about θ = 180ο.
RESONANCE DYNAMICS ... (GLADMAN/KAVELAARS)AA
more complicated), only the tracks in (p,θ) space; in fact the
separatrix motion requires a formally-infinite time.
ORBITAL RESONANCE DYNAMICS
In celestial mechanics the ‘resonant angle’ θ is formed from a
combination of angles which describe the orbit and positions
along the orbit of the two particles. For example, for describing Pluto’s motion in the 3:2 resonance, the resonant angle
θ32 = 3λ − 2λ N − ϖ
(7)
λ = ϖ+ M
(8)
where
is the mean longitude of the particle, made up of the longitude
of pericenter ϖ = Ω + ω of the orbit and the mean anomaly M
of the particle. The mean anomaly M of celestial mechanics is
the angular position of the particle past the pericenter point, but
in the ‘average’ sense rather the strict geometrical angle; the
interested reader should refer to an orbital dynamics text for a
deeper understanding, but here it is sufficient to think of λ as
the angle that starts from the arbitrary reference direction from
which the angular position of the ascending node Ω is measured (where the particle crosses the reference plane going
north) around to the particle (this is precisely true for circular
orbits in the reference plane).
The resonant angle θ32 is not a simple angle to interpret; one
cannot draw it in coordinate space as an angle terminating at
the plutino particle. This angle appears in the expansion of the
‘disturbing function’ of celestial mechanics, where it causes a
singularity in the theory if the particle is precisely at the resonant semimajor axis. However, analytical treatments (beyond
the scope of this article) can be developed. Some geometrical
insight is obtained if one asks the question of the value of θ32
at an instant when the plutino is at perihelion. In such a case
M = 0 for the particle (since M is measured from pericenter)
and Eq. 8 indicates λ = ϖ (that is, the angle around the particle
from the reference direction is the same as the angle to perihelion, which must be true of course), so that
θ32 = 3ϖ − 2λ N − ϖ = 2(ϖ − λ N ) .
(9)
180o,
then this equation
If the resonant angle has the value of
demands that the perihelion longitude of the TNO is ϖ = λN +
90o, meaning that the perihelion point is 90o ahead of
Neptune’s position. Because of the 360o degeneracy of angles,
another valid possibility is that ϖ = λN B 90o, corresponding to
perihelion 90o behind Neptune. Thus, even a plutino with an
orbit so eccentric that its closest solar approach is nearer the
Sun than Neptune’s distance of 30 AU (Pluto and many other
plutinos have such large eccentricities) is protected from close
encounters with Neptune by the resonance condition. If the resonant angle is near, but not exactly, 180o, a significant offset is
still induced between the perihelion longitude and Neptune’s
location.
Because of the gravitational tugs which Neptune exerts on the
plutino, the heliocentric orbit of the small body precesses and
the perihelion direction
relative to Neptune
would drift. One would
expect that eventually
the perihelion direction
of the plutino could
align with Neptune. The
reader should confirm
that although both the
longitudes of the plutino
and Neptune advance
rapidly, because of the
fact that Neptune’s period is 2/3 of the plutino’s, Eq. 7 then results
in dθ32/dt being much
smaller than either Fig. 4 The history of the resonant
argument θ32 over 10 million
d λ/dt
or
d λN/dt.
years for a so-called ‘plutino’
Looking at Fig. 3, the
in the 3:2 mean-motion resoresonant particle’s tranance with Neptune. On time
jectory follows one of
scales of 104 years the resothe dashed curves inside
nant argument librates around
the separatrix, and it is
180o with an amplitude of
easy to see that as one
about 40o.
follows such a trajectory
away from θ = 180o
there will be an extremal value of θ at which point the angle
begins to return toward θ = 180o; the angluar deviation of this
extremal value is called the ‘libration amplitude’ of the trajectory. This libration of the resonant argument θ is the effect of
the resonance, and it manifests itself in configuration space as
a correlation between the plutino’s perihelion direction and that
of Neptune. (It is also true that at conjunction, where λ = λN
and thus Neptune and the plutino ‘line up’, the plutino is forced
to be near its maximum distance from the Sun.) Fig. 4 shows a
numerically-integrated time history of θ for a plutino with a
libration amplitude of about 40o. The θ32 resonant argument
completes one libration in only about 20 thousand years, but
this libration is stable for time scales of billions of years. The
small irregularities in Fig. 4 are caused by beating of the sampling interval and the small effects of the other planets.
The libration’s effect in space is illustrated in Fig. 5, which
requires some explanation. This is for the same orbit numerically integrated in the previous figure. Imagine looking down
on the Solar System from the north ecliptic pole (that is, directly above the plane of the Solar System), with the Sun at the
coordinate-system origin. Fig. 5 shows the location of Neptune
(N) this year, at a distance of 30 AU from the Sun (and nearly
circular orbit); Neptune would trace out a counter-clockwise
circle of this radius in an intertial frame, with a circle of that
radius shown for reference. However, this figure is drawn in a
reference frame that co-rotates with Neptune so that the planet’s location remains fixed at its current position. In this corotating reference frame plutinos execute a complex motion
like a child’s spirograph, going clockwise around the Sun
except for a small ‘loop’ near each pericenter which, as predicted above, occur near the points 90o ahead and behind Neptune.
Two orbits of the
plutino around the
Sun correspond to
one complete clockwise rotation around
this figure (thus with
two perihelion passages of course). The
orbit’s
pericenter
location then wobbles
slowly back and forth,
with the wobble
direction reversing at
the extremal points of
the resonant argument.
Fig. 5
The motion of a plutino in the
reference frame co-rotating with
Neptune. The two circles mark
heliocentric distances of 30 and
40 AU. Neptune’s position in
indicated by the large “N”. See
text for discussion.
REAL
PLUTINOS
The
phase-space
structure of resonances presented in
the previous section
makes clear the special dynamics of
objects in these resonances, and how the resonant configuration can influence the dynamics. However, except for Trojan
asteroids, there were few real-world examples to provide case
studies to compare to.
The 1930 discovery of Pluto created a puzzle regarding the origin of the orbit of this unusual object. The orbit is highlyinclined to the plane of the Solar System (i ~ 17o), a semimajor axis of ~ 39.7 AU, and with a large-enough eccentricity
(e ~ 0.25) that Pluto crosses the orbit of Neptune! These unusual orbital characteristics appeared to indicate that Pluto’s orbit
might not be stable.
Determining that Pluto and Neptune have orbital periods that
are near the ratio of 3:2 merely requires a simple application of
Kepler’s Third Law (Eqn. 1). Simply having commensurate
orbital periods, however, is not a sufficient condition to ensuring that Pluto and Neptune are in resonant orbits because the
combination of the angular variables must be in the right range
and the resonant argument must librate. The answer regarding
the stability of Pluto’s orbit needed to wait until 1965, at which
time the computational power became available to conduct a
numerical integration of Pluto’s orbital evolution under the
influence of the Sun and the four giant planets.
Cohen and Hubbard [1] demonstrated, via orbital integration,
that the 3:2 resonant angle between of Pluto with respect to
Neptune (θ32 as expressed in Eqn. 7) librates around a mean
80o and period of
value of 180o with an amplitude of
~ 19670 years. This libration of the resonant angle proved that
the orbits of Pluto and Neptune are in resonance, but better
observational data was required to secure the exact value of the
libration amplitude. This yin-yang between observational dis-
210 C PHYSICS
IN
covery and computational modelling is a key feature of Kuiper
belt research which continues to this day. At the time of writing, analytical/numerical understanding of resonance dynamics
and migrational capture into resonances (see below) is slightly
ahead of the observational data available.
A number of investigators have pursued ever-more detailed
investigations of Pluto’s orbital evolution, revealing that Pluto
is simultaneously trapped in a number of other types of resonances inside the 3:2 mean-motion resonance. Discussion of
the full complexity of Pluto’s orbit is beyond the scope of this
article and the interested reader is encourage to examine the
review by Malhotra and Williams [2].
Soon after the discovery of other Kuiper Belt objects in the
1990s, other trans-neptunian objects were realized to be
trapped in the 3:2 resonance. These objects were coined plutinos in analogy with Pluto. Other mean-motion resonances were
then shown to be inhabited as well; Ref. [3] reviews the early
development of the knowledge of the Kuiper Belt’s resonant
structure.
How do objects with an orbit like Pluto end up in this and other
resonances? Formation of Pluto in a Neptune-crossing orbit
seems unlikely. To understand the complexity of the problem
requires a consideration of the process of planet formation and
the general reversibility of Newtonian dynamics.
RESONANCE CAPTURE
Models of planetary accretion suggest that planets grow from a
smooth disk of material initially via coagulation of dust particles which form into cm-sized dust balls and then coalesce to
form larger like Pluto (see Ref. [4] for a review of the core
accretion model of planet formation). Of critical importance
for the current discussion is the finding that encounters
between growing planetesimals must have very low relative
velocities, or else the encounters are disruptive. For nearby
orbits of small eccentricity and mutual inclination, encounter
velocities between objects scale as vk e 2 + i 2 , where νk is the
keplerian orbital speed of one of the objects at the encounter.
Thus, although an object with a low-inclination, low-eccentricity orbit could perhaps form in resonance with Neptune it
seems highly unlikely that an object on a highly inclined or
eccentric orbit, like Pluto, could have done so.
Based on expectations surrounding the planet formation
process, Pluto most likely formed in a nearly circular orbit that
was tightly confined to the ecliptic plane. After formation,
some event(s) or action(s) must have provided the gravitational excitation required to leave Pluto in a high-inclination/higheccentricity orbit. How can such a process have occurred and
left Pluto in resonance with Neptune? Explaining the coupling
between Pluto’s orbital resonance in the context of the formation of the outer Solar System is where orbital dynamics provides its great constraint on models of planet formation.
During the final stages of the giant planet formation, and
Neptune in particular, there was likely a surviving disk of at
least a few Earth-masses of planetesimals orbiting in the eclip-
tic. When these particles are gravitationally scattered by the
giant planets, an exchange of angular momentum occurs. If the
planetesimal scatters inward of the giant planet then the giant
planet’s orbit will grow slightly or if the planetesimal scatters
outward then the giant planet’s orbit will shrink. The exact outcome, for the giant planet, depends on the net effect of all these
scattering events. For each outward scattering the change of
semi-major axis is approximately:
δa
m
a MN
where m is the mass of the scattered planetesimals and MN is
that of Neptune [5].
In the case of the outer Solar System, outward scattering off
Neptune rarely results in the planetesimal’s ejection out of the
Solar System while inward scatter usually leads to encounters
which hand the planetesimals, via Uranus and Neptune, down
to Jupiter. Mighty Jupiter easily ejects the planetesimals out of
the system, and thus moves in. In essence this makes Neptune’s
total number of inward scattering events more numerous than
the outward scattering ones and so Neptune’s orbit slowly
drifts outward, growing in semi-major axis. The slow expansion of the outer Solar System via this planetesimal scattering
appears to be an inescapable result during the late stages of
giant planet formation.
Imagine now that Pluto has formed and is far beyond premigration Neptune, say just outside the location of Neptune’s
3:2 resonance at the time. As Neptune migrates outward and its
semi-major axis increases, the heliocentric distance corresponding to the 3:2 resonance also slowly moves outward, eventually reaching the point where it sweeps past Pluto’s location.
The dynamics of resonant capture are rather complex, although
there are well-known examples. Likely the most familiar is that
of the Moon’s tidal locking to the same period as its revolution
around the Earth. In that case a slowly-acting force (the tidal
dissipation acting on the Moon’s surface) de-spun the Moon’s
rotation until it was caputred into resonance and found an equilibrium.
In the study of capture into Neptune’s 3:2 resonance, the transition depends on the initial conditions (namely the eccentricity of Pluto’s non-resonant orbit), the rate of evolution and
eccentricity of Neptune’s orbit and the relative importance of
other non-resonant gravitational perturbations (such as the
graviational forces from the other planets). In essence, the
timescale for migration of Neptune must be long compared to
the timescale of the resonant perturbations (~ 104 yr). One can
heuristically think about resonance capture by examining
Fig. 3 and imagining that the pre-capture trajectory of the plutino is circulating on a trajectory near the origin. These curves
are drawn in the case of a fixed orbit for Neptune. As Neptune
migrates, the plutino orbit’s action (to use the hamiltionian terminology), which here is the radius I, increases and the particle will approach the separatrix. One can productively think of
this as the migratory motion actually ‘breaking open’ the separatrix near the cusp point along the θ = 0 axis, and when the
particle’s orbit reaches this point it might end up escaping into
the outer circulatory region (in which case the particle has
passed through the resonance without capture), or becoming
trapped into the librational region and thus captured into the
resonance. In the language of nonlinear dynamics, one says
that some of the initial conditions are part of the ‘basin of
attraction’ of the librational fixed point at θ = 180o (this analogy is only good if one imagines the planet migrating forever).
When the planetary migration ceases, the structure of the resonance locks onto the phase space diagram like Fig. 3 and particles remain on the trajectory that they find themselves.
A phenomenon not obvious from the above discussion is that
after capture, continued outward migration causes the expansion of the orbital eccentricity. Via angular-momentum conservations one can determine the growth of the eccentricity as a
function of the change in semimajor axis of the migrating
Neptune. Malhotra [5] found that
e 2final
1 aN , final
2
+ ln
ecapt
3 aN ,capt
(10)
where capt subscripts refers to the values at the instant of the
resonant capture, and N subscripts refer to Neptune rather
than Pluto. Thus, assuming that the pre-migration Pluto
had einitial . 0 and given the current eccentricity of Pluto,
(efinal ~ 0.25) Eq. 10 provides an estimate of the migration distance:
Δa = (aN , final − aN ,capt ) = aN , final × ⎡⎣1 − exp( −3 ∗ e 2final ) ⎤⎦
5 AU .
This assumes that Pluto was ‘just exterior to’ the 3:2 resonance
when Neptune started to migrate (minimizing the migration
distance) and that Pluto’s pre-capture eccentricity was 0 (maximizing the migration). Since the publication of this theory [5],
a large number of plutinos have been discovered and the median eccentricity of this population appears to be ~ 0.18 [6], so
explaining the plutinos eccentricities with migration would
require about half of the plutinos to be captured during the last
3 AU of Neptune’s migration. Neptune is acting somewhat like
a ‘snowplow’ (a good canadian analogy), with the plutinos first
captured finishing ‘highest up’ in eccentricity. If one posits that
all plutinos had initial e . 0 then the largest-observed e = 0.33
requires Neptune’s total migration to have been 8 AU.
However, some pre-migration stirring of the Kuiper Belt may
have occured, in which case the total outward movement of
Neptune would have been less.
Capture Efficiency
The dynamical process of resonance capture depends on a
number of physical parameters, including the migration rate
and eccentricity of Neptune and the particle at the time the resonance is crossed. Thus, we might hope to use the current number and orbital distribution of objects trapped in the 3:2 resonance, versus some estimate of the pre-migration distribution,
as an archaeological indicator of Neptune’s migration rate.
Since Neptune’s ancient migration rate is intimately tied to the
primordial density of planetesimals in the outer Solar System
we have, in effect, a measure of the density of material in the
solar system at the epoch of planet formation.
A number of authors have examined the various interplays
between the initial conditions of a disk of material beyond
Neptune and the rate of capture into resonance (see Refs. [711] for examples). The dependencies between capture efficiency and migration rate (Ref. [9], in particular) and the initial
eccentricity of captured material (e.g., Ref. [10]) have been
examined as well as the long-term stability of captured
objects [8].
The underlying driver for these studies has been an attempt to
reconcile the observed orbital distribution with the migration
theory. If the planetesimal disk beyond Neptune (the Kuiper
belt) was completely quiescent (< e > ~ < sin(i) > n 0.01), as is
required for planetesimal accretion to be effective, then migration capture would have been nearly 100% effective and the
majority of Kuiper belt objects should now be members of the
3:2 resonance. In addition, the currently-observed inclination
distribution among the plutinos extends to higher inclination
than the inclination produced via resonance migration.
Levison et al. [11] have
proposed that instead
of migration into a
pre-existing low-eccentricity Kuiper Belt,
Neptune, Uranus and a
large planetesimal population were all flung
outwards from initial
locations interior to
30 AU. Migration still
occurs, but during an
epoch of large eccentricity for Neptune
which greatly enhances
resonance
capture.
(Neptune’s eccentricity
subsequently drops to
the present value of
nearly zero). This
model better explains
some of the observed
features of the Kuiper
Belt’s orbital distribution.
TWOTINOS
The resonant dynamics
in the outer Solar
System are very rich.
Many of the resonances
are significantly more
complicated than the
simple picture developed above. Fig. 6
212 C PHYSICS
IN
Fig. 6
shows the same polar diagram as before, but now for the 2:1
mean-motion resonance with Neptune; trans-neptunian objects
at this distance can be seen in Fig. 1 with a 47.4 AU. The resonant argument is θ21 = 2λ B λN B ϖ, and if this argument
librates the object is resonant and (whimsically) called a ‘twotino’. The separatrix between the outer and inner circulation
regions and the resonant libration region looks similar to
before, but now the resonant region is broken up into 3 regions
by an additional separatrix, called the symmetric and asymmetric islands as defined in Fig. 6’s caption. Two types of resonant
motion now exist: (1) symmetric libration where θ21 oscillates
around 180o, as before, but as the figure shows this must occur
with large amplitude, or (2) asymmetric libration where the
libration center is not 180o, but rather one of two other possible values (the libration center depends on the eccentricity) and
occuring with a smaller possible libration amplitude.
Twotinos librating around all three libration centers are known
(Fig. 7). In fact, due to the gravitational forces from the other
planets, a given twotino can switch between symmetric and
The θ21 resonant argument trajectories for the
2:1 mean-motion resonance. Here the particle
eccentricity is the radial coordinate.
Separatrices are shown as heavy curves. There
are now 3 resonance regions. The larger symmetric region encloses θ = 180ο and allows
large libration amplitudes. There are also the
two small asymmetric libration islands, so
named because a trajectory following one of
them will never pass through 180o, but will
instead oscillate with small amplitude around
another average value (of roughly θ21 =120 or
240 degrees here). Note that there are no smallamplitude symmetric librators. For particles of
different eccentricities the angle defining the
libration center of the asymmetric islands
changes. Figure provided by A. Morbidelli.
Fig. 7
The distribution of libration centers and resonance amplitudes for objects securely known to
be in the 2:1 mean-motion resonance with
Neptune. The error bars for each object represent the current range of allowable solutions. As
expected, all the symmetric librators (which
will have libration centers at 180o) have large
libration amplitudes, and the smaller-amplitude
symmetric librators have libration centers at
depend on their eccentricity. The greater number of Kuiper Belt objects known with libration
centers near 90o compared to near 270o is likely an observational bias due to objects in the
270o asymmetric island spending much of their
time in the direction of the galactic center.
asymmetric libration over time. This more complex resonance
structure opens new windows into the past, because assuming
capture into the 2:1 resonance occured due to migration, the
multiple libration islands may not be equally populated.
Asymmetric Capture
The unusual orbit of Pluto and the explanation of that orbit as
the result of resonance capture during Neptune migration leads
naturally to the question of what other signatures of migration
might exist among the resonant Kuiper belt objects.
When a low-eccentricity object is initially captured into the 2:1
resonance, the effect of continued migration will be to cause
the newly-captured twotino’s eccentricity to grow, slowly leading it towards the separatrix entirely inside the resonant region
(shown in Fig. 6), forcing a choice between one of the two
asymmetric islands and the large libration-amplitude symmetric island. Which island will be selected?
As in the previous discussion, Fig. 6 is only true for the instantaneous, i.e. non-migrating, case. Ref. [12] numerically investigated the effects of migration-induced resonance capture on
the distribution of libration amplitudes and libration centers of
the twotino population. Remarkably, they found that the likelihood of capture into the so-called leading (θ < 180) and trailing (θ > 180) asymmetric libration islands is dependent on the
rate of Neptune’s ancient migration.
Murray-Clay and Chiang [12] further investigated the dynamics
and provide an explanation of the resultant asymmetry. In the
simplest case, a migrating Neptune causes the center of symmetric libration to shift slightly such that the average value of
θ is greater than 180o. Thus, as an object librates about the
shifted ‘symmetric’ resonance center, the object will spend
slightly more time on the side of the libration potential than
on the other side (see Fig. 8). This shift in libration center causes the asymmetry in the capture. Murray-Clay and Chiang further recognized that the libration amplitude is a function of the
object’s e prior to its capture into the 2:1 resonance. As the
libration amplitude increases, the amount of time an object
spends visiting phases space accessible to both asymmetric
islands grows. In fact, for very small libration only the ‘trailing’ island is explored while for slightly larger eccentricities
both islands can be explored.
The strength of preference for the ‘trailing’ versus the ‘leading’
asymmetric islands is caused by the size of offset in the center
of the symmetric libration. The size of this offset is, itself, a
function of the migration rate and initial eccentricty of the captured twotino (see equation 26 of Ref. [12]). Thus, measuring
the current ratio of ‘trailing’ to ‘leading’ twotinos could, when
coupled to some estimate of the initial eccentricity of the twotino population, provide another measure of the migration rate of
Neptune. The density of the primordial outer Solar System
could then be inferred.
Although both the 3:2 capture efficiency and the ‘trailing’ versus ‘leading’ 2:1 asymmetry provide only indirect, and somewhat complicated, information on the rate of migration of
Neptune, they both provide independent inferences. Coupling
of the observational census of the Kuiper Belt’s resonant populations with our evolving knowledge of the complex dynamical evolution of these bodies provides a path to unlocking the
ancient history of our Solar System.
OBSERVATIONS
Since the discovery of the second trans-Neptunian object in
1992 there has been a veritable explosion of discovery. At the
close of 2007, just 15 years after that transformative discovery,
approximately 807 trans-Neptunian objects
were known, of which some 120 might be
plutinos and another dozen or so appear to
be Twotinos (see Ref. [13] for a review of
the nomenclature of the trans-Neptunian
region and a census of its members). But
wait, why ‘approximately’ 807 and ‘might
be plutinos’ or ‘appear to be twotinos’? The
endeavour of wide-field search surveys for
transneptunian bodies has been an exciting
race, with multiple groups attempting to
establish strong records of discovering new
objects. Orbital determination of the discoveries, however, tends to be less exciting
and requires many times the observational
Fig. 8 The evolution of the libration angle (θ = θ21) with a migrating Neptune for two resources as compared to the initial discovtransneptunian objects which become captured in the 2:1 resonance. At t = 0 the object
ery. Determining a transneptunian’s precise
is circulating (not in resonance) and is captured into resonance near t ~ 0.09 Myr. The
orbit, particularly one in resonance,
left panel shows the low-eccentricity case where the libration amplitude is small and
the separatrix (dashed line) between the two asymmetric islands is offset to θ < π ; requires many dozen positional measureclearly the object spends no time with small θ and so can not be captured into the island ments spread over a number of years.
with θ < π. For objects with larger initial eccentricities (right panel) the libration ampli- Without a well-sampled and lengthy obsertudes are larger and the offset of the separatrix is smaller. In this case the object spends vational history, it is impossible to consome time on the small- side of the separatrix point and so capture into that island strain quantities like the libration center
becomes possible. (Figure provided by R. Murray-Clay and E. Chiang).
and amplitudes (see Fig. 7), or even
answering the question: “Is this object in resonance?”
The difficulty in determining the precise orbits leads to the
problem of object loss. Without secure estimates of the orbital
elements, the future positional uncertainty on the sky grows
rapidly and soon (within a few months to a year) observers can
no longer locate the object and it is lost. The constant leaking
away of these discoveries contributes to a biased view of the
known populations.
Determining the migration rate of Neptune from the relative
strengths of the population of resonant objects (such as the relative size of the population of the 3:2 resonance as compared
to say, the 2:1 resonant objects) requires that we either have
unbiased estimates of those populations or, if the estimate is
biased that we be able to account for that bias (see Ref. [14] for
a discussion of observational biases that must be accounted
for). Unfortunately, accounting for objects that have been lost
due to insufficient observations is not a bias that we can corrected for.
One observational bias that can be understood in a straightforward way is the flux bias: brighter objects are easier to detect.
Fig. 5 shows the positions, relative to the Sun in a frame rotating with Neptune, that a plutino explores during one libration
cycle; this object makes closer approaches to the Sun when
roughly 90 degrees away from Neptune on the sky. Since one
only detects trans-Neptunians via reflected sunlight, their
brightness L is given by L % 1/r4, where r is their heliocentric
distance. Thus, since plutinos are closest when 90 degrees from
Neptune, a survey that looks in these directions is more likely
to find plutinos than would a survey which looks towards
Neptune. Because of this flux bias, we must know the pointing
history of a survey before we determine the fraction of objects
in the 3:2 resonance compared to other populations.
Fig. 7 presents the libration centers and amplitudes for the current sample of twotinos. We see that most of the currently-
observed asymmetric population is in the ‘leading’ asymmetric
island (libration center, θ ~ 90o, see Fig. 6). Does this then indicate that Neptune’s migration was extremely slow or rather that
the pre-capture orbits of the twotinos had large eccentricities;
both effects would reduce the dynamical preference for ‘leading’ capture (see Ref. [12]) or is some other bias perhaps at
work? For twotinos, as for plutinos, there is a bias towards discovery when the objects are near perihelion, for small-libration
amplitude objects this occurs when they are near their libration
centers relative to Neptune. Currently the trailing libration center, which is 60o ‘behind’ Neptune, is aligned on the sky with
the direction of the galactic center. Thus, although the trailing
twotinos are brightest when on this part of their obit, discovery
is hampered by observational confusion with the vast number
of galactic plane stars in the background. Determining the
intrinsic population from the observed one requires a knowledge of the pointing history of the discovery survey and an
accurate estimate of the fraction of the search fields that are
actually discoverable.
The authors of this manuscript are involved in a project (visit
http://www.cfeps.net) that is striving to address the major problems with the observational constraints of Kuiper belt populations. This project carefully measures the internal biases for
discovery and tracking, so that the sample of objects can be
used to ‘back out’ the true populations.
CONCLUSION
Although this article has just scratched the surface of a very
complex dynamical problem, we hope the reader will have garnered some appreciation for the dynamics of resonant orbital
motion and how it can be used to diagnose the ancient history
of the Solar System.
ACKNOWLEDGEMENTS
We thank A. Morbidelli and E. Chiang for helping provide
modified figures.
REFERENCES
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
Cohen, C.J. and Hubbard, E.C., Astronomical Journal, 70, 10 (1965).
Malhotra, R. and Williams, J.G., in Pluto and Charon, Edited by S. Alan Stern, and David J. Tholen, University of Arizona Press, p.
127, (1997).
Davies, J.K., McFarland, J., Bailey, M.E., Marsden, B.G., and Ip, W.-H., in The Solar System Beyond Neptune, eds. M.A. Barucci, H.
Boehnhardt, D.P. Cruikshank, and A. Morbidelli, University of Arizona Press, Tucson, p. 11-23 (2008).
Lissauer, Jack J., Annual Review of Astronomy and Astrophysics, 31, 129, (1993).
Malhotra, R., Nature, 365, 819, (1993).
Kavelaars, Jj, Jones, L., Gladman, B., Parker, J.W., and Petit, J.-M., Astronomical Journal, submitted (2008b).
Malhotra, R., Astronomical Journal, 110, 420, (1995).
Gomes, R., Nature, 426, 393, (2003).
Chiang, E.I. and Jordan, A.B., Astronomical Journal, 124, 3430, (2002).
Hahn, J.M. and Malhotra, R., Astronomical Journal, 130, 2392, (2005).
Levison, H.F., Morbidelli, A., Vanlaerhoven, C., Gomes, R., and Tsiganis, K., Icarus, 196, 258, (2008).
Murray-Clay, R.A. and Chiang, E.I., Astrophysical Journal, 619, 623, (2005).
Gladman, B., Marsden, B.G. and Vanlaerhoven, C., in The Solar System Beyond Neptune, eds. M.A. Barucci, H. Boehnhardt, D.P.
Cruikshank, and A. Morbidelli, University of Arizona Press, Tucson, p. 43 - 57 (2008).
Kavelaars, Jj, Jones, L., Gladman, B., Parker, J.W., and Petit, J.-M., in The Solar System Beyond Neptune, eds. M.A. Barucci, H.
Boehnhardt, D.P. Cruikshank, and A. Morbidelli, University of Arizona Press, Tucson, p. 59- 69, (2008).
214 C PHYSICS
IN
ARTICLE DE FOND
VISUALIZING THE INVISIBLE
OBSERVATIONS
USING
POLARISATION
JO-ANNE C. BROWN, JEROEN M. STIL, AND TOM L. LANDECKER
BY
I
t has been known for centuries that the Earth has a
magnetic field. The idea that the Galaxy has a magnetic field and that that field might play an important
role in the physics of the Galaxy is more recent, dating back almost 60 years. Fermi [1] proposed that cosmic
rays may be generated outside the solar system as a result
of sufficiently energetic particles colliding with moving
irregularities in an interstellar magnetic field. In his view,
this magnetic field would not only be a generator of cosmic rays, but also a containment factor to prevent the rays
from escaping the Galaxy. Indeed, it is now believed that
magnetic fields and cosmic rays contribute to the vertical
support of the gas in the Galaxy [2].
In addition to playing a significant role in pressure balance, magnetic fields play an essential role in star formation, by inhibiting gravitational collapse of interstellar
clouds – primary star formation regions – and by remov-
SUMMARY
An electromagnetic wave can be uniquely
characterized by the four Stokes parameters:
I, Q, U, and V. Typical observations in astronomy rely solely on total intensity measurements of the incoming radiation (Stokes I).
However, a significant amount of information
both about the emitting region and the propagation path is carried in the remaining
Stokes parameters. These data provide a
means to observe parts of the interstellar
medium which remain invisible in Stokes I, at
any wavelength. For example, when an electromagnetic wave propagates through a
region containing free electrons and a magnetic field, the plane of polarisation of the
wave will rotate - an effect recorded only in
Stokes Q and U. The interstellar medium of
the Galaxy is such a region, containing free
electrons (observed as HII) and a magnetic
field of a few microgauss. By imaging in
Stokes Q and U we are able to observe signatures of magnetic field perturbations from
the small scale (tens of pc) to the large scale
(kpc). In this paper, we review the status of
Canadian polarisation studies of cosmic
magnetic fields and discuss the leading role
Canada is playing in polarisation studies
around the world.
ing prestellar angular momentum [3]. Consequently, magnetic fields directly affect the distribution of stars. It is also
believed that magnetic fields influence galaxy formation
and evolution by causing large density fluctuations which
result in structures within a galaxy [4].
In situ measurements of interstellar magnetic fields are not
yet possible. Even the NASA Voyager Space probes,
launched in 1977, have only just reached the
Heliosheath [5], and are not expected to reach the interstellar medium (ISM) until 2013 at earliest. Furthermore, signatures of interstellar fields are only indirectly observable
through polarisation observations. The majority of astronomical observations are done in total intensity, much like
that of a standard photograph, thus rendering magnetic
fields ‘invisible’ in most astronomical data. Consequently,
magnetic fields have been either largely ignored in astronomy or have been used as a scape-goat for otherwise unexplained phenomena.
It has only been in the last 30 years that magnetic field
observations outside our local arm have made advances,
with some of the most significant steps being led by
Canadians and Canadian instrumentation. For example, in
almost stereotypical form, a little Canadian facility, built
and run on a shoe-string budget (by any ‘large-scale’ facility standard) revolutionized polarisation observations with
a method that would be emulated worldwide. That little
facility is the Synthesis Telescope of the Dominion Radio
Astrophysical Observatory (DRAO), operated by the
Herzberg Institute for Astrophysics of the National
Research Council. The array, built amid the mountains of
the Okanagan Valley, in south-central British Columbia,
consists of seven antennas in a linear east-west configuration. Three of the antennas rest on railway tracks, allowing
for variable baselines (ie. pairs of correlating antennas),
while the remaining four are fixed. The antennas were salvaged from various locations: two were surplus moonradar antennas, two were surplus troposcatter antennas,
two came from Five Colleges Radio Astronomy
Observatory in Massachusetts, and one came from Texas
where it had been used for solar radio astronomy. In order
to meet weight-limit requirements at the focus, the feed
horns were fashioned at a lampshade factory and the
waveguides were constructed of irrigation piping. To pressurise the feed horns with dry air, aquarium pumps and
canning jars are used. Yet, with clever engineering, the telescope is able to obtain high sensitivities and resolution,
producing some of the best radio images in the world [6].
J.C. Brown <jocat@
ras.ucalgary.ca>,
J.M. Stil, Centre for
Radio Astronomy,
University of Calgary,
Calgary, AB T2N 1N4
and
T.L. Landecker,
Dominion Radio
Astrophysical
Observatory, National
Research Council
Canada, Penticton,
B.C. V2A 6J9
VISUALIZING THE INVISIBLE ... (BROWN ET AL.)
This observatory was the primary instrument of the Canadian
Galactic Plane Survey (CGPS [7]), a multi-facility project
designed to image a section of the Galactic plane at multiple
wavelengths and full polarisation. After demonstrated success
in both observation technique and discoveries, the CGPS was
then expanded into the International Galactic Plane Survey,
which included the Southern Galactic Plane Survey
(SGPS [8,9]) with observations taken from the Australia
Telecope Compact Array (ATCA). The survey areas for both
projects are shown in Figure 1. While there have been many
significant discoveries made through these surveys, the ones
we will focus on in this paper relate to polarisation observations in general, and magnetic fields in particular. We will also
discuss how our knowledge gained with these projects has prepared us, as Canadians, for the international stage with large
upcoming projects like the Square Kilometer Array.
Polarisation and Stokes Parameters
The concepts of polarisation and Stokes parameters are often
minimized or bypassed altogether in the undergraduate curriculum in physics, though they are of fundamental importance in
electrical engineering, where radio astronomy has its roots. In
fact, all man-made electromagnetic signals are polarised
because antennas are made of wires which channel the electron
flow, imposing a preferred direction on the emitted radiation.
The only signals that can be unpolarised are natural signals.
We briefly review the relevant aspects of polarisation and the
Stokes parameters in this section, and discuss how they are
exploited to remotely study magnetic fields in the subsequent
sections. Additional details may be found in Refs [11] and [12].
Electromagnetic waves are transverse, meaning that their oscillations are perpendicular to their direction of propagation. If
the wave vector and wave electric field define a plane that does
not change as the wave propagates, then the wave is linearly
polarised, since the wave is seen to define a line when viewed
along the direction of propagation.
If we consider two orthogonal, linearly polarised electromagnetic waves of the same frequency travelling in the z direction,
with the first polarised in the x direction (x-z plane), and the
second in the y direction (y-z plane), the electric fields of the
two waves may be described by the following equations:
Fig. 1
All sky view (in Galactic coordinates) of past, present, and
future radio observation surveys with significant Canadian
involvement. CGPS: Canadian Galactic Plane Survey (PI at
UofC); SGPS: Southern Galactic Plane Survey (UofC participation); VLA: Very Large Array observations (Co-PI at
UofC) ; GALFACTS: Galactic Arecibo L-band Feed Array
Continuum Transit Survey (PI at UofC); ASKAP: Australian
Square Kilometer Array Pathfinder (Canada is a formally recognized partner). The Global Magneto-Ionic Medium Survey
(GMIMS: PI at NRC) will cover the entire sky.
TECHNIQUES FOR OBSERVING MAGNETIC
FIELDS
The interstellar medium consists of several basic constituents:
atomic, molecular and ionized gas, dust, cosmic rays and magnetic fields [10]. Most of the constituents have some form of
observable radiation, and may be observed directly at the
appropriate wavelength. Such observations are referred to as
‘total intensity’ or ‘Stokes I’ observations. Unlike these other
constituents, magnetic fields themselves do not radiate, and
consequently, cannot be observed directly. However, they can
affect the sources of radiation or the radiation directly, given
the right conditions. The signature of these effects shows up in
the other Stokes parameters, primarily U and Q, as we discuss
below.
Ex = E1 cos (kz − ωt)
(1)
Ey = E2 cos (kz − ωt − δ)
(2)
where k is the wavenumber, ω is the frequency, t is time, and δ
is the phase offset between the two waves (δ = δx − δy ). The
detected wave will be the vector sum of these two individual
waves such that E = Ex^i + Ey ^j. At z = 0, the components of E
may be reduced to:
Ex = E1 cos (ωt)
(3)
Ey = E2 cos (ωt + δ)
(4)
Combining equation 3 and equation 4 results in the equation of
an ellipse:
1 = aEx2 − bExEy + cEy2
where
a=
1
2 cos δ
1
, c= 2 2 .
, b=
E12 sin 2δ
E1 E2 sin 2 δ
E2 sin δ
IN
(6)
This equation describes the locus of points traced out by the
vector E as it propagates. The ellipse, known as the polarisation ellipse, may be characterized by two angles, τ and ε, as
illustrated in Figure 2.
Critical to our work is the angle τ, known as the polarisation
angle. It gives a measure of inclination of the ellipse with
respect to the x axis 1 and is defined within the limits of
1. In astronomy, the x axis is defined as ‘sky North’.
216 C PHYSICS
(5)
VISUALIZING THE INVISIBLE ... (BROWN ET AL.)AA
0 o # τ < 1 8 0 o,
since τ = 180o is
indistinguishable
from τ = 0o. The
second angle, ε , is
in essence a measure of the ellipticity
of the wave. Its
value is determined
by the arc-cotangent
of the ratio of the
major axis (OA)
Fig. 2 The polarisation ellipse (after fig- to the minor axis
ure 4.5 in Ref. [11]). E1 and E2 are (OB) and is defined
the magnitudes of two monochro- within the limits
matic waves of identical frequency, − 45o # ε # + 45o.
polarised in the x and y directions Negative values of ε
respectively. The vector sum of correspond to ‘rightthese two waves traces out an handed’ (or rightellipse with a polarisation angle of
elliptically
τ (see text for more details).
polarised) waves,
where E moves
counterclockwise when viewed travelling towards the observer, while positive values of ε correspond to ‘left-handed’ (or
left-elliptically polarised) waves, where E moves clockwise as
viewed from the same vantage point. This definition of handedness is the IEEE convention. It is the standard in radio
astronomy and is consistent with the well known ‘right hand
rule’. 2
Depending on the properties of E1, E2 and δ, the polarisation
ellipse will take on different forms. In general, waves with
0o < δ < 180o will be left elliptically polarised whereas waves
with 180o < δ < 360o (−180o < δ < 0o) will be right elliptically
polarised. If δ = 0o or δ = 180o , the wave will be linearly
polarised. In cases where E1, E2 and δ are such that elliptical
polarisation results, the wave can be thought of as having some
component of linear polarisation, and some component of circular polarisation.
The state of polarisation represented by a polarisation ellipse
may be described mathematically by the four Stokes parameters. Introduced in 1852, they are
I = ( E 21 + E 22 )/Z
Q=
(E 21
−
E 22 )/Z
(7)
= I cos 2ε cos 2τ
U = (2E1E2 cos δ)/Z = I cos 2ε sin 2τ
V = (2E1E2 sin δ)/Z = I sin 2ε
(8)
(9)
(10)
where Z is the impedance of the medium [11]. Stokes I is the
total intensity of the wave, Stokes Q and U are measures of the
linear polarisation of the wave, and Stokes V is a measure of
the circular polarisation of the wave.
With these two formulations of the Stokes parameters (the relationship between the two may be found in Ref. [11]), it is easy
2. Under the classical physics convention, the handedness definition is reversed.
to see that the first definition allows for straight-forward measurements of the parameters by an antenna with a given impedance Z designed to measure linear polarisation on two orthogonal axes. Once the Stokes parameters have been determined,
the second formulation allows for the calculation of τ and ε. In
particular, we note that
τ=
U
1
tan −1
Q
2
(11)
The above discussion dealt with a completely polarised or
monochromatic wave, where E1, E2 and δ are constant. In general, emissions from celestial radio sources extend over a wide
range of frequencies. Within any finite range of frequencies
detected by a receiver, the wave will consist of a superposition
of a large number of statistically independent waves with a
variety of polarisations. Therefore, E1, E2 and δ will be detected as having time dependence, and the Stokes parameters will
use the time-averages of these values.
In the pure, monochromatic case, I 2 = Q 2 + U 2 + V 2. With
multiple wave fronts averaged together, it is possible to have
I 2 $ Q 2 + U 2 + V 2. For a completely unpolarised wave, Q =
U = V = 0. The degree or fraction of polarisation is defined as
dp =
Q2 + U 2 + V 2
polarised power
=
total power
I
(12)
where 0 # dp # 1. Thus, dp = 1 for a completely polarised
wave, while dp = 0 for a completely unpolarised wave.
In the interstellar medium, most of the polarised radiation we
observe at radio wavelengths comes from synchrotron emission (see section below), and is consequently linearly polarised.
Therefore, when we talk about polarised intensity, we are really talking about linearly polarised intensity, defined as
PI = U 2 + Q 2 .
(13)
Figure 3 shows CGPS Stokes I, linear polarised intensity (PI)
and polarisation angle (τ) images of a small part of the Galactic
plane. The CGPS images are unique because of their image
fidelity, and the inclusion of short spacing information over a
large area of the Galactic plane. The most striking aspect of
Figure 3 is the wealth of structure in polarisation angle and
intensity, which is not seen in Stokes I. In fact, it is rare to find
a counterpart of polarised structure in the total intensity
images. An exception is the low polarised intensity in the upper
right corner, associated with depolarisation by tangled magnetic fields in a low-density halo around the bright HII (ionized
hydrogen) region W4 seen in the Stokes I image [13] .
Faraday Rotation
When a linearly polarised electromagnetic wave propagates
through a region of free electrons permeated by a magnetic
field (e.g. a magnetized plasma such as the interstellar medium), its plane of polarisation will rotate. This phenomenon is
known as Faraday rotation. Faraday rotation can be understood
as a consequence of birefringence, where the magnetised plas-
and if
Ω
sec θ
ω
1
(14)
ω 2p
(15)
1,
ω2
where Ω is the cyclotron frequency, and ωp is the
plasma frequency. If the total strength of the Galactic
magnetic field is, on average, roughly Bo = 10 μG
(10−9 T), the electron cyclotron frequency is
Ω=
eB
me
(16)
= 175.63 rad/s
where e and me are the electron charge and mass,
1
respectively. With (Ω/ω) sec θ = 1000
, and ω = 2π H
1420 MHz, the QL approximation holds for θ <
89.99887o. Similarly, for equation 15, using ne =
1 cm -3 (106 m-3), the plasma frequency is
⎛ n e2 ⎞
ωp = ⎜ e ⎟
⎝ ε me ⎠
Fig. 3
Mosaic MX1 of the CGPS in total intensity (Stokes I ), total linear polarised
intensity (with and without the zero-spacing data), and polarisation angle.
“Zero spacings” or single-dish data have been added to the interferometric
data for all images except the ‘interferometer only’ polarised intensity. For
the purpose of this illustration, the polarisation angle has been shifted by
90o. As shown, the polarisation angles have a relatively narrow distribution,
which is a consequence of the local large-scale magnetic field.This plot
demonstrates how different a region can look in polarisation compared to
Stokes I, as well how the addition of zero-spacing data can affect the polarisation images themselves. W4 and W5 are ionized hydrogen (HII) regions
first identified by Westerhout [14]. The box highlights the Faraday rotation
feature first identified by Gray et al. [15].
ma has two different indices of refraction corresponding to two
different states of incident polarisation [16].
1
2
(17)
= 56 H 103 rad/s
As a result, ω2p /ω2 = 4 x 10−11 n 1. The fact that the
QL approximation holds for such a large range of
angles at radio wavelengths often leads one to forget
that it is an approximation that must be verified
depending on the application and region of the ISM
being explored.
Invoking the QL approximation and using the resultant indices of refraction for circularly polarised
waves in the ISM plasma (see Ref. [10]), the amount
of rotation a radio wave will acquire, Ψ, is given by
Ψ = λ2 (0.812IneBAdl ) [rad]
= λ2 RM
(18)
For a linearly polarised wave, the birefringence is with respect
to right and left circularly polarised waves (an alternative to the
linearly polarised basis set used above). The birefringence will
slow one of the circularly polarised waves with respect to the
other, resulting in a rotation of their sum, the linearly polarised
wave (e.g. Ref. [17]).
where λ is the wavelength in units of m, ne is the electron density in units of cm−3, B is the magnetic field in units of μG, dl
is the incremental pathlength in units of pc, and RM is the rotation measure:
To calculate the indices of refraction for a cosmic magnetised
plasma, the quasi-longitudinal (QL) approximation is
invoked [18]. The validity of the QL approximation depends on
^ ) coinhow closely the direction of propagation of a wave (k
^@k
^) ,
cides with the field direction (^b ), defined as θ = cos−1(b
and on the electron density and the collision frequency. As θ
approaches 90o a linearly polarised wave will aquire some
ellipticity (known as the Cotton-Mouton effect), instead of simply rotating as it would in the QL regime.
It is important to recognize the significance of three key elements of equation 18. First, it is wavelength dependent. As a
result, waves of different frequency will experience different
amounts of rotation through the same plasma. Second, the
effect of Faraday rotation is weighted by the electron density;
higher electron densities will result in greater rotation. Finally,
it is the direction of the line-of-sight component of the magnetic field (B||) that determines the sign of the rotation measure.
Since the path length is defined to be from the source to the
receiver, (ie. the telescope on Earth), a magnetic field with B||
directed towards us results in a positive RM, while a magnetic
field with B|| directed away from us results in a negative RM.
The QL approximation is valid if [19]
218 C PHYSICS
IN
RM = 0.812 IneB A dl [rad m−2]
(19)
With this in mind, if we assume that at all wavelengths, the
polarised emission from a given source is emitted at the same
polarisation angle, τo, and that the radiation is only affected by
Faraday rotation, then the detected polarisation angle, τ at a
given wavelength λ, will be given by
τ = τo + λ2 RM.
Radio emission at decimeter wavelengths from
our Galaxy is dominated
by synchrotron emission
from relativistic electrons with Lorentz factor
Γ >∼ 105 in a magnetic
field. Relativistic particles from interstellar
space, named cosmic
rays for historic reasons,
were first observed on
Earth in balloon experiments by Hess in 1912.
(20)
Since this relationship is linear, measurements of τ at multiple
wavelengths can determine the RM for a given source as the
slope of the graph of τ versus λ2.
The ease with which RMs can be determined, coupled with the
significance of the sign of the RM, makes RM measurement a
powerful tool for probing the ISM magnetic field. Pulsars and
extragalactic sources (EGS) are sources of linearly polarised
radiation and are often used as compact (or point source)
probes of the Galactic magnetic field. Since it is possible to
estimate the distance to these sources, and if we know something of the electron density along their lines-of-sight, then we
can work backwards to estimate what the magnetic field must
look like along their particular plumb-lines. Subsequently, the
goal for observations is to measure RMs for the highest density of sources possible, allowing for the most accurate reconstruction of the intervening field.
Prior to the CGPS, observations of multiple polarisation angles
for EGS were done at widely separated wavelengths, often at
different times, and sometimes even at different facilities.
Consequently, there was uncertainty as how to ‘unwrap’ the
polarisation angles in order to determine the correct RM (e.g.
Refs. [20,21]). DRAO was the first facility to do polarisation
measurements at 4 wavelengths sufficiently close together so
that the ambiguity in RM calculations was removed [22]. This
technique was emulated at the ATCA for the SGPS. Instead of
4 bands, the SGPS had 12 bands, improving on the technique
initiated by the CGPS.
Not only did the CGPS set the standard for observation techniques, it also set the standard for EGS RM source density.
Prior to the CGPS, Broten et al. [23] had compiled a catalog of
high quality EGS RM measurements. With 674 sources in the
catalog, the majority of which were out of the Galactic plane,
the RM density was roughly 1 source per 60 square degrees.
The CGPS produced (and continues to produce) RMs at a density of 1 source per square degree, resulting in significantly
more reliable conclusions about the magnetic field than previously possible. The SGPS source density is slightly lower than
the CGPS, at 1 source per 2 square degrees, as a result of depolarisation through the inner Galaxy. However, it must be noted
that prior to the survey, there was only 1 published EGS RM in
the entire SGPS region.
Synchrotron emission
The classic source of radiation is accelerating charges.
Therefore, charged particles moving in the presence of a magnetic field will undergo acceleration through the Lorentz force,
and subsequently radiate. If the particles are moving at relativistic speeds, this radiation is called synchrotron radiation.
Fig. 4
Diagram of a relativistic electron
with Lorentz factor Γ spiraling in
a uniform magnetic field. The
electron emits synchrotron radiation with a high degree of linear
polarisation in a narrow cone
directed along the instantaneous
velocity of the electron.
Figure 4 shows a relativistic electron on a helical path in a uniform
magnetic field due to the
Lorentz force. In the
observer’s rest frame, the
radiation emitted by the
accelerated electron is
emitted in a narrow cone along the electron’s velocity vector as
a result of the relativistic beaming effect (e.g. Ref. [24]). The
opening angle of the emission cone is 1/Γ, much narrower than
shown in Figure 4. Consequently, the observer sees only emission from those electrons that have an instantaneous velocity
directed towards the observer (pitch angle α). Furthermore,
electrons traveling parallel to the field will not be accelerated,
and will therefore not radiate. Thus, the amount of synchrotron
emission observed depends on the presence of a magnetic field
component perpendicular to the line of sight, Bz. Optically thin
synchrotron emission 0s at frequency ν of an ensemble of relativistic electrons with a power law energy spectrum N(E) ~ Eγ
in a uniform magnetic field depends on the magnetic field component in the plane of the sky Bz according to
0s ~ Bz(γ+1)/2ν−(γ −1)/2
(21)
Synchrotron emission provides information about the slope of
the energy spectrum of the electrons, and an estimate of the
strength of the magnetic field if assumptions are made about
the volume of the source and the energy density of cosmic rays.
The minimum combined energy density of cosmic ray particles
and magnetic field for an observed source is similar to the
equipartition energy density of the magnetic field. Magnetic
field estimates from the brightness of synchrotron emission
assume this minimum energy condition, without clear justification that the minimum energy density condition or equipartition apply.
Synchrotron emission from a region with a uniform magnetic
field has a theoretical limit of ~ 70% [25], with the plane of
polarisation perpendicular to the direction of Bz in the plane of
the sky. Polarisation of synchrotron emission therefore gives
information on the magnetic field component perpendicular to
the line of sight, while Faraday rotation gives information on
the magnetic field component along the line of sight.
In practice, the observed emission is integrated over large
regions in space with a complicated magnetic field structure
and an ionized plasma present in addition to the relativistic
electrons. At decimetre wavelengths the integrated emission is
usually much less polarised than the theoretical 70%. In this
wavelength range thermal emission is mostly fainter than synchrotron emission unless the line of sight crosses a denser HII
region, ionized by massive stars. Nevertheless, it is very common for a plasma structure to cause significant Faraday rotation, yet to be undetectable by its thermal emission. Faraday
rotation alone cannot alter the amplitude of the polarised signal, but a number of physical and instrumental effects often
reduce the fractional polarisation of the recorded signal [26,27].
Differential Faraday rotation (or depth depolarisation) occurs
when synchrotron emission generated at different depths along
the line of sight suffers different rotation, and vector averaging
reduces the observable polarised intensity. Beam depolarisation occurs when many turbulent cells and/or large RM gradients exist within the beam of the telescope, again leading to
vector averaging. On most angular scales these effects create
structures in polarised intensity, and particularly in polarsation
angle, that have no counterpart in total intensity. Since these
effects are most pronounced at decimetre wavelengths, that
wavelength regime provides the best data for studying the magnetic field configuration within the interstellar medium.
The detection of linear polarisation in the extended Galactic
radio emission [28,29] provided crucial evidence in establishing
the synchrotron mechanism that makes the Milky Way a strong
radio source. The apparent potential of polarisation observations to reveal the Galactic magnetic field led to efforts to map
polarised emission over wide areas of the sky. The best of these
datasets is that of Brouw and Spoelstra [30] who presented data
from the Dwingeloo 25-m Telescope for much of the Northern
sky at four frequencies between 408 and 1411 MHz. Angular
resolution ranged from 2o to 36N, but the sampling was far from
complete. A major Canadian contribution to this field is the
1.4 GHz polarisation survey of Wolleben et al. [31] made with
the DRAO 26-m Telescope. The northern sky was mapped
down to declination -30o with 200 times more data points than
the Dwingeloo data and five times better sensitivity, but based
on the absolute calibration of Brouw and Spoelstra. These new
data have played an important role in the cosmology industry,
by providing a significant counterpoint to the Wilkinson
Microwave Anisotropy Probe (WMAP) 23 GHz data [32] as
well as providing a clearer understanding of the polarised features observed (e.g. Ref. [33]).
MAGNETIC STRUCTURE IN THE LOCAL ISM
Cosmic ray electrons in the Galaxy emit synchrotron emission
that is observed as a featureless glow across the sky with a tendency to be brighter near the Galactic plane. In addition to this
smooth synchrotron background we see synchrotron emission
from specific objects, mainly supernova remnants. The diffuse
synchrotron emission originates from a large volume in space.
It takes a substantial line of sight distance to build up
detectable synchrotron emission, because the highly relativistic
220 C PHYSICS
IN
electrons that emit this radiation are so rare. The same volume
of space is littered with plasma structures that give rise to
Faraday rotation and depolarisation effects described above. It
comes as no surprise that the interpretation of polarisation of
diffuse Galactic emission is very difficult. However, it is also
the only way we can observe most of the magneto-ionized
interstellar medium.
Wieringa et al. [34] reported structures in polarised radio emission at 327 MHz that had no counterpart in total intensity. It
was soon realized that these structures were small-scale modulations in the polarisation angle of Galactic synchrotron emission that were detectable by the radio interferometer. The structure in polarisation angle gave rise to structure in the Stokes Q
and U images, even though the interferometer did not detect the
smooth emission in total intensity because of the so-called
missing short spacings; Faraday rotation effects tend to break
large structures into structure on smaller angular scales.
Nevertheless, inclusion of data from a large single-dish radio
telescope provides information on the largest structures that an
interferometer cannot detect and has a dramatic effect on the
polarisation images. Some polarised features are observed to
change significantly with the inclusion of the large-scale
polarised emission not seen by the interferometer (see
Figure 3).
The study of structure in diffuse polarised emission really took
off with the CGPS. The polarisation survey of the CGPS at
1.4 GHz [35] marks a major advance in polarization observations. Data from the DRAO Synthesis Telescope, the
Effelsberg 100-m Telscope, and the DRAO 26-m Telescope
have been combined to give accurate representation of all
structures down to the resolution limit of ~1N. With 1.7 H 107
independent data points, this is the largest polarisation survey
made to date, and the most extensive dataset to combine singleantenna and aperture-synthesis data. For this survey, new techniques were developed to correct for instrumental polarisation
across the field of view [36,37] and independent calibrations of
the three datasets were carefully compared.
An origin in Faraday rotation means automatically that the
polarised “objects” will usually not look like things seen in
other wavelengths. Since the polarised sky is almost entirely
new, the first task is to classify the various polarisation features, though it is often difficult to draw a boundary around
such objects. Early results from the survey have identified two
enigmatic Faraday rotation features [15,38], with a similar object
reported by Haverkom et al. [39]. Polarisation features associated with known objects have also been identified, including that
associated with a planetary nebula [40] and a stellar wind bubble [41].
Figure 3 shows the lenticular feature of Gray et al. [15] which
resides in the direction of the HII region W5. The longest size
of this feature is approximately two times the diameter of the
full moon. While visible in polarisation angle, it was not visible in polsarised intensity prior to the addition of the zero-spacings, nor is there any obvious counterpart in Stokes I. While the
origin of such “polarisation lenses” is still not clear, it is likely
maintained? We cannot hope to
answer these questions until we
understand what it really looks
like: where it is (and isn’t!)
located, what its direction and
magnitude are, and how it is
correlated (or uncorrelated)
with the medium in which it
appears to be embedded.
At one time, it was believed
that the Galactic magnetic field
was primordial in origin, meaning it was present as a weak
‘protofield’ at the time the
Fig. 5 (a) Number density of polarised sources in the NVSS from Stil and Taylor [42]. Contours are drawn Galaxy was formed, and it subat 2.7 and 4.0 sources per square degree. The black area in the lower left corner is below the southsequently evolved and ampliern declination limit of the NVSS. (b) Hα intensity in the same region from the SHASSA survey [43].
fied as the protogalaxy contracted and rotated. The prieither a concentration of electron density or a magnetic field
mary objection to the primordial theory is based on the time
structure.
scales required to generate the observed fields existing in
galaxies [47].
Looking at the polarisation images of the CGPS in general, it
is almost impossible to find the Galactic plane: the polarised
However, a primordial field may have served as the seed field
signals seem to continue undimmed to the southern and northfor a Galactic dynamo [48]. By definition, a dynamo converts
ern limits of the data. This is in sharp distinction to the totalthe energy of motion of a conductor into the energy of an elecintensity distribution and the distributions of other ISM tracers
tric current and a magnetic field [49]. In the Galaxy, the consuch as dust. The simplest interpretation is that the polarised
ducting fluid requirement for a dynamo is satisfied by the ionemission that we are seeing at 1.4 GHz is generated in nearby
ized interstellar gas. The differential rotation of the Galaxy
volumes of the ISM. This is consistent with the concept of the
could produce appropriate fluid motions that would amplify the
polarisation horizon introduced by Uyaniker et al. [44]. The
seed field. Dynamo theory is currently favored among magnetcombined effects of depth depolarisation and beam depolarisaic field theorists as it appears to be quite robust and seems to be
tion do not allow us to detect polarised emission beyond a cerable to provide a universal explanation of the varied field contain distance. The distance to the polarisation horizon depends
figurations observed [50].
on frequency, beamwidth, and direction.
This is where much of the observational work is focussed Some local structures escape detection in observations of difidentifying the topology of the field to provide adequate confuse polarised emission. Figure 5 shows a large magnetized
straints for modeling in order to determine the most likely
shell in the Gum nebula, that was revealed because it depolarismode(s) of the dynamo(s) acting in the Galaxy. While some
es extragalactic radio sources [42]. The shell is so large on the
features of the Galactic magnetic field are universally accepted
sky that the Big Dipper would fit inside it. Figure 5(a) shows
as facts, others remain highly contentious. We discuss both the
the sky density of polarised extragalactic sources found in the
accepted and contentious features of the Galactic magnetic
NRAO VLA Sky Survey (NVSS; Ref. [45]). The dark ring repfield in the following sections.
resents a lack of polarised sources, depolarised because of
Accepted Observational Constraints
strong Faraday rotation in a shell with a compressed magnetic
field. The upper part of this shell is visible in the Hα emission
The Galactic magnetic field is usually considered to be comin Figure 5(b). The lower part of the shell is clearly defined by
posed of two distinct components: a smooth or uniform comthe lack of polarised sources in Figure 5(a), but bright Hα
ponent, Bu, with scale sizes on the order of a few kiloparsecs
emission that does not depolarise background sources confus(kpc), and a turbulent or random component, Br, with scale
es the image in Figure 5(b). Vallée and Bignell [46] first sugsizes on the order of tens of parsecs (pc; Ref. [51]). The unigested the presence of a magnetised shell because of anomform component is observed to be concentrated in the
alously high rotation measures for some extragalactic sources.
disk [48,52], with a dominant azimuthal component, some radial
component (indicating a spiral field), and a weak vertical or z
THE GLOBAL MAGNETIC FIELD IN OUR
component. Conversely, Br is believed to be isotropically distributed as integrated along the line-of-sight [53], though there
GALAXY
is some evidence to suggest that Br is correlated with Bu [54],
Despite its recognized importance, very little is truly known
and that the scale-sizes of Br are significantly different
about the Galactic magnetic field. What is its source? How is it
between and within the spiral arms [55,56].
necessary to explain the observations, thus coming full circle to
the initial conclusion of no reversals beyond the solar circle.
Fig. 6
The directions of the large-scale Galactic magnetic field as
viewed from the North Galactic Pole. The grey scale is the
CL02 model of the electron density [57]. Q1-4 indicate the four
Galactic quadrants, and the asterisk indicates the location of
the Sun. Solid arrows indicate universally accepted field
directions; single-ended dashed arrows indicate field directions as supported by Canadian work, though not necessarily
universally accepted; double-ended arrows indicate regions
remaining highly debated with no recent Canadian input.
As shown in Figure 6, the field within the local arm is observed
to be directed clockwise, as viewed from the North Galactic
pole, with a strength of roughly 6 μG [53,58] 3. In the first quadrant (Q1) of the Sagittarius-Carina arm, the magnetic field is
unquestionably observed to be directed counter-clockwise [59,60], indicating a region of magnetic shear between the
local and Sagittarius-Carina arms. Such a region is what we
call a magnetic field reversal. The number and location of such
magnetic field reversals are arguably the most significant factors in differentiating between likely dynamo modes.
Controversial Observational Constraints
The paucity of available data makes identifying magnetic field
reversals exceedingly difficult. Consequently, different interpretations of similar data do occur.
Based primarily on RMs of EGS, neither Simard-Normandin
and Kronberg [52] nor Vallée [61] could find any evidence for
reversals beyond the solar circle. However, using a limited
number of pulsar RMs along with the EGS RMs available at
the time, Rand and Kulkarni [62] and Clegg et al. [63] suggested
the presence of a field reversal associated with the Perseus arm,
and Han et al. [64] concluded that there may be an additional
reversal beyond the Perseus arm. With the significant increase
in EGS RM source density of the CGPS, coupled with newer
pulsar RMs, Brown et al. [22] demonstrated that a reversal is not
3. For perspective, the strength of the Earth’s magnetic field is 0.6 G at the poles.
222 C PHYSICS
IN
In the inner Galaxy, studies using both pulsars and EGS RMs
have produced evidence suggesting the field reverses back to a
clockwise direction at R ~ 5.5 kpc for the Scutum-Crux
arm [21], and perhaps switches again at R ~ 3 kpc for the
Norma arm [64]. The apparent pattern of the field reversing with
every arm is supported by the recent work of Weisberg et
al. [65], who studied pulsar RMs primarily in Q1. Han et al. [66]
has even suggested the field reverses at every arm-interam
interface. Conversely, using the new SGPS EGS RMs along
with the pulsar RMs, Brown et al. [67] could only find strong
evidence for one reversal, and weaker evidence for a second
inside the solar circle. Furthermore, the strong first reversal is
seen to occur between the Sagittarius-Carina arm and ScutumCrux arm in quadrant 4 (Q4), instead of between the local arm
and Sagittarius-Carina arm as observed in Q1. This suggests
the field has much less inclination (ie. it is more azimuthal)
than the optical spiral arms. This is in agreement with the interpretation of the pulsar RM data made by Vallée [68] who envisaged a ring model with a reversal that passes through the
Sagittarius-Carina arm around Galactic longitude = 0 (ie. the
Sun - Galactic centre line).
Clearly more data are required to differentiate between these
differing opinions. To that end, we recently acquired time on
the VLA to fill in the gaps between the CGPS and the SGPS,
as shown in Figure 1. We hope these data will be sufficient to
determine the field structure in the inner Galaxy. Otherwise, we
will have to wait for data from the upcoming projects described
in the section on Current and Future Projects with Significant
Canadian Content.
THE MAGNETIC FIELD IN EXTERNAL SPIRAL
GALAXIES
While living inside of a galaxy provides unique opportunities
to study galactic dynamics up close, it carries with it the inherent problem of the ‘forest-for-the-trees’ effect. Therefore, it is
extremely beneficial if we balance observations within our own
Galaxy with those of external spiral galaxies.
Observing magnetic fields in external galaxies has the advantage that it is easier to see structure from the outside, and magnetic fields can be studied in a variety of galaxies with different properties. However, different techniques must be
employed than for observations of the Galactic magnetic field.
Even for galaxies as close as a few Mpc, present-day telescopes
cannot detect a sufficient number of background EGS to do statistical RM studies as is done in our Galaxy. To date, only two
external galaxies, M31 [69] and the Large Magellanic
Cloud [70], have been probed with RMs of background sources.
Both galaxies show a large-scale, regular magnetic field, but no
field reversals were detected.
Most observations of magnetic fields in external galaxies are
limited to polarisation of synchrotron emission with a resolution up to ~10NN, corresponding with 240 pc at a distance of
have a magnetic field oriented along the bar. Some galaxies
with an intense star burst and an associated outflow of gas perpendicular to the disk display large-scale magnetic fields perpendicular to the disk, and far into the halo, e.g. NGC 4569 [73].
Gravitational interaction with another galaxy or ram pressure
interaction with an intracluster medium can deform the disk of
a galaxy and its associated magnetic field [74].
Fig. 7
Polarised intensity (contours) of the nearby (5.5 Mpc) spiral
galaxy NGC 6946 on a gray scale image of Hα emission [72].
Our line of sight is almost perpendicular to the plane of the
disk of NGC 6946. Vectors indicate the magnetic field direction derived by rotating the observed plane of polarisation by
90o. Radio emission from the inter-arm regions in this galaxy
is ~ 30% - 40% polarised, indicating a very regular magnetic
field. Figure kindly made available by R. Beck.
5 Mpc. All but the largest polarised structures identified in diffuse Galactic emission (section on Magnetic Structure in the
Local ISM) would be unresolved in radio images of nearby spiral galaxies. As the resolution is also similar to the outer scale
of energy injection into the interstellar medium by stars, most
structures in the interstellar medium only contribute to an unresolved stochastic component of the rotation measure. This
structure within the beam leads to depolarisation of the emission. Remarkably, radio emission of some galaxies is locally
30 - 40% polarised at a wavelength of 6 cm [71,72] .
Images of polarised emission of spiral galaxies reveal the magnetic field on a galactic scale, projected on the plane of the sky.
As in our own galaxy, the regular magnetic field is predominantly in the azimuthal direction in the plane of the disk. The
direction of the regular magnetic field shows a spiral pattern
similar to the optical spiral arms, but the most regular fields are
found away from the spiral arms in inter-arm regions. Figure 7
shows polarised emission of the galaxy NGC 6946 [72] in relation to the spiral arms as traced by the Hα emission of massive
star formation regions. The total magnetic field in the (optical)
spiral arms is actually a factor ~2 stronger than in the highly
polarised magnetic arms, but the field in the spiral arms is more
tangled on scales smaller than the resolution of the image,
resulting in a low degree of polarisation.
Large-scale departures from a symmetric, azimuthally oriented
magnetic fields are found in some galaxies. Barred galaxies can
Observational evidence for the evolution of magnetic fields in
galaxies with cosmic time has been elusive because radio
observations of spiral galaxies at cosmologically interesting
distances are not possible with current instruments. Indirect
evidence for substantial magnetic fields in normal galaxies
comes from the association of high rotation measures of distant
quasars with absorption systems at lower redshift. Bernet et
al. [75] and Kronberg et al. [76] found that quasars with optical
Mg II absorption systems with a redshift smaller than that of
the quasar, indicative of an unrelated normal galaxy in the line
of sight to the quasar, have a substantially higher spread in rotation measure than quasars without Mg II absorption. The
inferred magnetic field strengths are similar to those in nearby
spiral galaxies, leading Kronberg et al. [76] to the conclusion
that magnetic field strengths similar to those in present day
galaxies already existed a few Gyr after the big bang. The first
predictions of the contribution of spiral galaxies to deep
polarised radio source counts were made by Stil et al. [77].
CURRENT AND FUTURE PROJECTS WITH
SIGNIFICANT CANADIAN CONTENT
Canada has gained a reputation for excellence in sensitive wide
field polarisation imaging. Continuing work on the DRAO
deep fields [78] provides the most sensitive polarisation image
of the sky to date. Canada also participates in a number of
international projects that are bound to revolutionize our understanding of the origin and evolution of cosmic magnetic fields.
For much of the future work on Galactic magnetism, the significant quantity is not only polarisation angle, but specifically
rotation measure. All projects listed below will utilize multiple
channels allowing for studies in rotation measure synthesis,
where images may be formed at individual values of RM [79].
The Global Magneto-Ionic Medium Survey (GMIMS; principal investigator M. Wolleben, NRC) is an international project
utilizing several facilities from around the world to map
polarised emission across the entire sky from 16 cm to 1 m
(300 MHz to 1.8 GHz). Specialized receivers designed and
built at DRAO are being used by the participating telescopes.
With observations commencing in April of 2008, the project
will survey the diffuse polarised emission from the Galactic
disk to the halo, at a resolution of 0.5 degrees.
Complementary to GMIMS is the Galactic Arecibo L-band
Feed Array Continuum Transit Survey (GALFACTS; principal
investigator A. R. Taylor, Calgary). GALFACTS is a polarisation survey with the Arecibo radio telescope that will have a
sensitivity of μJy and hundreds of spectral channels. A new
multi-beam cleaning technique was developed in Calgary to
make high-fidelity images of compact polarised sources and
diffuse emission with the seven-beam Arecibo L-Band Feed
Array (ALFA). GALFACTS will extend to 32% of the sky the
kind of analysis that has so far been restricted to small deep
fields. The many frequency channels and large bandwidth will
open the possibility to study the wavelength-dependent polarisation in much more detail than any previous survey.
The Very large Array (VLA) in New Mexico is currently being
upgraded to become more than an order of magnitude more
sensitive than before. Key to the upgrade of the VLA is the new
central correlator that has been designed and built at DRAO.
The Expanded Very Large Array will be a much more versatile
instrument with a larger instantaneous bandwidth, suitable to
make deep polarisation images of the sky.
Canada is a partner in the Australian Square Kilometre Array
Pathfinder (ASKAP [80]). This technology demonstrator for the
much larger Square Kilometre Array, scheduled to be completed by 2020, will explore wide-field imaging (30 square degree
instantaneous field of view) at bandwidth of 300 MHz divided
into 16000 frequency channels. Apart from the engineering
challenges for this new-generation radio telescope, SKA
pathfinders such as ASKAP provide new challenges in terms of
image calibration and processing. Canada is expected to play a
leadership role in developing techniques for wide-field polarisation imaging and calibration for both SKA pathfinders
(including ASKAP) and the SKA itself. Finally, one of the five
science drivers for the SKA is the origin and evolution of cosmic magnetism. Canada has demonstrated expertise on both
the technical and scientific fronts defining the SKA, and we
will undoubtedly continue to do so.
ACKNOWLEDGEMENTS
The authors acknowledge support from the Natural Sciences
and Engineering Research Council through the Discovery
Grants program. The Canadian Galactic Plane Survey is a
Canadian project with international partners and is also supported by the Natural Sciences and Engineering Research
Council. The Dominion Radio Astrophysical Observatory is
operated as a national facility by the National Research
Council.
REFERENCES
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
19.
20.
21.
22.
23.
24.
25.
26.
27.
28.
29.
30.
31.
32.
33.
34.
35.
E. Fermi, Physical Review, 75, 1169, (1949).
A. Boulares, and D.P. Cox, ApJ, 365, 544, (1990).
E.G. Zweibel & C. Heiles, Nature, 385, 131, (1997).
E. Kim, A.V. Olinto, & R. Rosner, ApJ, 468, 28, (1996).
E.C. Stone, A.C. Cummings, F.B. McDonald, B.C. Heikkila, N. Lal, N., & W.R. Webber, Nature, 454, 71, (2008).
T.L. Landecker, et al., A&AS, 145, 509, (2000).
A.R. Taylor, S.J. Gibson, M. Peracaula, et al., AJ, 125, 3145, (2003).
N.M. McClure-Griffiths, J.M. Dickey, B.M. Gaensler, A.J. Green, M. Haverkorn, & S. Strasser, ApJS, 158, 178, (2005).
M. Haverkorn, B.M. Gaensler, N.M. McClure-Griffiths, J.M. Dickey, & A.J. Green, ApJS, 167, 230, (2006).
L. Spitzer, Jr., Physical Processes in the Interstellar Medium, John Wiley & Sons, Inc., (1978).
J.D. Kraus, Radio Astronomy, Cygnus-Quasar Books, Powell, Ohio, (1986).
J.C. Brown, “The Magnetic Field in the Outer Galaxy”, PhD Thesis, University of Calgary, (2002)
A.D. Gray, T.L. Landecker, P.E. Dewdney, A.R. Taylor, A.G. Willis, & M. Normandeau, ApJ, 514, 221, (1999).
G. Westerhout, Bull. Astron. Inst. Netherlands, 14, 215, (1958).
A.D. Gray, T.L. Landecker, P.E. Dewdney, & A.R. Taylor, Nature, 393, 660 (1998).
E. Hecht, Optics, Addison Wesley Longman Inc., (2001)
F.F. Chen, Introduction to Plasma Physics and Controlled Fusion, Plenum Press, (1984).
J.A. Ratcliffe, 1962 The Magneto-Ionic Theory and Its Applications to the Ionosphere, Cambridge University Press, (1962)
I.H. Hutchinson, Principles of Plasma Diagnostics, Cambridge University Press, (1987)
M. Simard-Normandin, P.P. Kronberg, & S. Button, ApJS, 45, 97, (1981).
R.J. Rand, & A.G. Lyne. MNRAS, 268, 497 (1994).
J.C. Brown, A.R. Taylor, R. Wielebinski, & P. Mueller, ApJL, 592, L29, (2003).
N.W. Broten, J.M. MacLeod, & J.P. Vallée, ApSS, 141, 303, (1988).
D.J. Griffiths, Introduction to Electrodynamics, Prentice-Hall, Inc., (1999).
A.G. Pacholczyk, Radio Astrophysics, Series of Books in Astronomy and Astrophysics, San Francisco: Freeman (1970).
B.J. Burn, MNRAS, 133, 67, (1966).
D.D. Sokoloff, A.A. Bykov, A. Shu kurov, E.M. Berkhuijsen, R. Beck, & A.D. Poezd, MNRAS, 299, 189, (1998).
G. Westerhout, C.L. Seeger, W.N. Brouw, & J. Tinbergen, Bull. Astron. Inst. Netherlands, 16, 187, (1962).
R. Wielebinski, J.R. Shakeshaft, & I.I.K. Pauliny-Toth, “The Observatory”, 82, 158, (1962).
W.N. Brouw, & T.A.T. Spoelstra, A&AS, 26, 129, (1976).
M. Wolleben, T.L. Landecker, W. Reich, & R. Wielebinski, A&A, 448, 411, (2006).
L. Page et al., ApJS, 170, 335, (2007).
M. Wolleben, ApJ, 664, 349, (2007).
M. Wieringa, A.G. de Bruyn, D. Jansen, W.N. Brouw, & P. Katgert, A&A, 268, 215 , (1993)
T.L. Landecker et al., in preparation (2008)
224 C PHYSICS
IN
36.
37.
38.
39.
40.
41.
42.
43.
44.
45.
46.
47.
48.
49.
50.
51.
52.
53.
54.
55.
56.
57.
58.
59.
60.
61.
62.
63.
64.
65.
66.
67.
68.
69.
70.
71.
72.
73.
74.
75.
76.
77.
78.
79.
80.
T. Ng, T.L. Landecker, F. Cazzolato et al., Radio Science, 40, 5014 , (2005).
R.I. Reid, A.D. Gray, T.L. Landecker, & A.G. Willis, Radio Science, 43, 2008 (2008).
B. Uyaniker, & T.L. Landecker, ApJ, 575, 225, (2002).
M. Haverkorn, P. Katgert, & A.G. de Bruyn, A&A, 356, L13, (2000).
R. Ransom, B. Uyaniker, R. Kothes, & T. Landecker, ApJ, 684, 1009, (2008).
R. Kothes et al., in preparation (2008)
J.M. Stil, & A.R. Taylor, ApJ, 663, L21, (2007)
J.E. Gaustad, P.R. McCullough, W. Rosing, & D. Van Buren, PASP, 113, 1326, (2001).
B. Uyaniker, T.L. Landecker, A.D. Gray, & R. Kothes, ApJ, 585, 785, (2003).
J.J. Condon, W.D. Cotton, Greisen et al., AJ, 115, 1693, (1998).
J.P. Vallée, & R.C. Bignell, ApJ, 272, 131, (1983).
E.N. Parker, ApJ, 401, 137, (1992).
J.L. Han, and G.J. Qiao, AAP, 288, 759, (1994).
E.N. Parker, Scientific American, 249, 44, (1983).
R. Beck, A. Brandenburg, D. Moss, A. Shukurov, D. Sokoloff, ARA&A, 34, 155, (1996).
A. Ruzmaikin, D. Sokolov, and A. Shukurov, Nature, 336, 341, (1988).
M. Simard-Normandin, & P.P. Kronberg, ApJ, 242, 74, (1980).
C. Heiles, in Polarimetry of the Interstellar Medium, W. Roberge & D. Whittet (eds), 97 (ASPCS), 457, (1996).
J.C. Brown, A.R. Taylor, ApJ, 563, L31, (2001).
M. Haverkorn, B.M. Gaensler, J.C. Brown, N.S. Bizunok, N.M. McClure-Griffiths, J.M. Dickey, & A.J. Green, ApJ, L33, (2006).
M. Haverkorn, J.C. Brown, B.M. Gaensler, & N.M. McClure-Griffiths, ApJ, 680, 362, (2008).
J.M. Cordes and T.J.W. Lazio, astro-ph/0207156 (2002).
R. Beck, SSRv, 99, 243, (2001).
M. Simard-Normandin, and P.P. Kronberg, Nature, 279, 115, (1979).
R.C. Thomson, & A.H. Nelson, MNRAS, 191, 863, (1980).
J.P. Vallée, A&A, 124, 147, (1983).
R.J. Rand, & S.R. Kulkarni, ApJ, 343, 760, (1989).
A.W. Clegg, J.M. Cordes, J.M. Simonetti, & S.R. Kulkarni, ApJ, 386, 143, (1992).
J.L. Han, R.N. Manchester, & G.J. Qiao, MNRAS, 306, 371, (1999).
J.M. Weisberg, J.M. Cordes, B. Kuan, K.E. Devine, J.T. Green, & D.C. Backer, ApJS, 150, 317, (2004).
J.L. Han, R.N. Manchester, A.G. Lyne, G.J. Qiao, & W. van Straten, ApJ, 642, 868, (2006).
J.C. Brown, M. Haverkorn, B.M. Gaensler, A.R. Taylor, N.S. Bizunok, N.M. McClure-Griffiths, J.M. Dickey, & A.J. Green, ApJ, 663,
258, (2007).
J.P. Vallée, ApJ, 619, 297, (2005).
J.L. Han, R. Beck, & E.M. Berkhuijsen, A&A, 335, 1117, (1998).
B. Gaensler, M. Haverkorn, L. Staveley-Smith, J.M. Dickey, N.M. McClure-Griffith, J.R. Dickel, & N.M. Wolleben, Science, 307,
1610, (2005).
E.M. Berkhuijsen, R. Beck, & P. Hoernes, A&A, 398, 937, (2003).
R. Beck, A&A, 470, 539, (2007).
K.T. Chyży, M. Soida, D.J. Bomans, B. Vollmer, Ch. Balkowski, R. Beck, & M. Urbanik, A&A, 465, 472, (2006).
M. Weżgowiec, M. Urbanik, B. Vollmer, R. Beck, K.T. Chy¿y, M. Soida, & Ch. Balkowski, A&A, 471, 93, (2007).
M.L. Bernet, F. Miniati, S.J. Lilly, P.P. Kronberg, & M. Dessauges-Zavadsky, Nature, 454, 302, (2008).
P.P. Kronberg, M.L. Bernet, F. Miniati, S.J. Lilly, M.B. Short, & D.M. Higdon, ApJ, 676, 70, (2008).
J.M. Stil, A.R. Taylor, M. Krause, & R. Beck, PoS(MRU)069, arXiv:0802.1374 (2007)
A.R. Taylor, J.M. Stil, J.K. Grant, T.L. Landecker, R. Kothes, R.I. Reid, A.D. Gray, D. Scott, P.G. Martin, A.I. Boothroyd, G. Joncas,
F.J. Lockman, J. English, A. Sajina, & J.R. Bond, ApJ, 666, 201, (2007).
M.A. Brentjens, & A.G. de Bruyn, A&A, 441, 1217, (2005).
S. Johnston et al., “Publications of the Astronomical Society of Australia”, 24, 174, (2007).
DEADLINE EXTENDED
= ?
TOO BORING?
TROP ENNUYEUX?
WIN CASH - GAGNER DE L’ARGENT
CONTEST FOR NEW LOGO IS UNDERWAY
DEADLINE EXTENDED T O JANUAR Y 23, 2009
The Editorial Board is inviting all CAP members, friends, or colleagues to submit designs
for a new PiC-PaC logo which should fit well in the upper left hand corner of the front
cover of each issue, and integrate well on any of the PiC covers.
The winning entry will be featured on the 2009 January-March issue and the photograph
and bio of the submitter will be published in the issue. The winner will receive $150, an
“Art of Physics” t-shirt and, if applicable, a one-year membership in the CAP.
UN CONCOURS POUR UN NOUVEAU LOGO EST EN COURS
DATE LIMITE PROLONGÉE JUSQU’AU 23 JANVIER 2009
Le Comité de rédaction invite tous les membres de l’ACP, amis, et collègues à soumettre
des croquis de logos PiC-PaC pour remplacer celui qui se trouve au coin en haut à
gauche de la couverture. Le logo devrait bien s’intégrer dans ce coin supérieur gauche
quelque soit la conception de la page couverture.
Le croquis choisi figurera dans le numéro de Jan/mars 2009 avec une photographie et une
courte biographie de l’artiste. Le gagnant recevra 150 $, un T-shirt « Art de la physique »
et si, pertinent, une adhésion d’un an à l’ACP.
226 C PHYSICS
IN
ARTICLE DE FOND
METAL-POOR STARS:
THE INTERSECTION OF CHEMISTRY, COSMOLOGY,
AND STARS
BY
KIM VENN
T
he chemical and kinematic differences between
the stellar subpopulations hold clues about how
the Milky Way formed and evolved. Stellar
motions can be scrambled by a variety of processes, such as close encounters with other objects, but this is
a much smaller effect compared to the original orbit of the
gas clouds when a particular star first formed and became
part of the Galaxy. The fact that the stars in the halo of the
MWG have random and eccentric orbits in all three directions has been interpreted as the halo component forming
first from a collapsing gas cloud. The remaining gas would
have then collapsed into a pancake, to conserve angular
momentum, and went on to form the thick and thin disk
components. The bulge is the center of the potential that
defines the Galaxy, thus it is expected and found to contain a bit of everything. This model for the formation of
the MWG is called monolithic collapse, where the Galaxy
formed from an enormous cloud of gas and in relative isolation, and was first described by Eggen, Lynden-Bell, and
Sandage [1]. While this simple model has successfully
explained many attributes of our Galaxy, it is no longer
championed.
Today, we live in a Universe that is dominated by Cold
Dark Matter (CDM), not to mention an unknown energy
source (dark energy, L) as shown by the cosmological
interpretations of the microwave background power spectrum (e.g., the recent 5-year WMAP results [2]). In a CDM
Universe, large structures like the MWG form through the
SUMMARY
Like other spiral galaxies, the Milky Way
Galaxy (MWG) has several distinct structural
components that probably appeared at different stages in its formation process. The
Sun is a part of the “thin disk”, but the MWG
has at least three other distinct stellar components; the “thick disk”, the “halo”, and the
“bulge”. The orbits and kinematics of the
stars and other objects in each component
are primarily what sets them apart from one
another. But there are other characteristics
that differ, such as the ages and the chemical
compositions as well.
hierarchical accretion of smaller structures, such as dwarf
galaxies [3]. Evidence for the effects of merging are seen
through tidal streams in ours and other spiral galaxies [4,5].
New galaxy formation computer simulations in a LCDM
Universe are also starting to be able to make realistic
galaxies, including the various distinct structural components [6,7].
But looks only get you so far. While the computer simulations are able to make realistic looking galaxies, the only
way to constrain these models and also to compare them
with the simpler monolithic collapse models are through
observational tests. One of the most useful tests of galaxy
formation is to examine the chemical contents of the
stars [8,9]. The chemical composition of a star’s outer layers is, for the most part, preserved from birth. Thus, since
some stars formed early during the formation of the MWG
and others have formed recently, with apparently continuous star formation at various rates throughout time, then
the amounts of various chemical elements and the build up
of the elements can be used, like fossils, to probe the formation and evolution of the MWG.
CHEMISTRY AND KINEMATICS OF STARS
To test models of galaxy formation requires the chemistry
and kinematics of a large number of stars in the MWG.
Unfortunately, most of the stars analysed to date are from
a small region very close to the Sun. Nevertheless, we are
able to sample all of the Galactic components since we can
determine the orbital kinematics of each star. In general,
stars in the MWG halo are metal-poor, followed by the
thick disk, the thin disk, and the most metal-rich stars are
found in the Galactic bulge [8,10]. In Figure 1, the kinematical assignments for stars with detailed abundance determinations are shown; bulge objects are not shown on these
plots since a slightly different assignment was used for
those.
Kim Venn
<[email protected]>,
Department of
Physics and
Astronomy, University
of Victoria, Victoria,
BC, V8P 5C2
The overall metallicity is interesting because most of the
chemical elements are made in stars, thus stars that are
more metal-poor are thought to have formed at the earliest
epochs, before a galaxy has undergone very much star formation or chemical evolution. Of course, this is only an
assumption since inhomogeneous chemical mixing could
leave pockets of pristine gas in a galaxy (or more likely, on
METAL POOR STARS ... (VENN)
stars in each of the Galactic components indicating an age progression.
Fig. 1
Kinematic diagram showing the Galactic rotational velocity
(V) as a function of a combination of the radial velocity into
the disk (U) and the perpendicular velocity out of the disk
(W) to make T = U 2 + W 2. The stars are allocated to Galactic
components using velocity ellipsoid probabilities; thin disk
(red), thick disk (green), halo (cyan). Two additional kinematic components stand out; an extreme retrograde component (black) and a highly elliptical orbital component (blue).
the outskirts of a galaxy); also it is possible that pristine gas
falls into a galaxy at a later time, locally diluting the interstellar gas. Thus, metal-poor does not necessarily mean earliest
epoch of formation, but it’s not a bad starting approximation.
Regardless, since the stars themselves are the nucleosynthetic
furnaces that make most of the chemical elements, then over
time the chemistry in a galaxy must evolve. In Figure 2, the
iron (Fe) abundances are shown relative to Galactic rotational
velocity, and while the scatter is quite large (and real) for halo
stars, in general there is an increasing trend in metallicity with
It is amazing that the detailed chemistry of an individual star
can be determined. Stars emit a continuous spectrum from their
hot interiors, but this travels through the cooler outer layers of
the stellar atmosphere before reaching us. Thus, we see an
absorption line spectrum from every star, which typically
includes strong lines of hydrogen or molecular bands, depending on the temperature of the stellar atmosphere. The strength
of any absorption lines depends on far more than just the abundance of the element in the stellar atmosphere though; it
depends critically on the temperature (thus affecting the atomic ionization and excitation states), and less so on the luminosity and the surface velocity fields (due to collisions which can
affect line strengths), and on the stratification of these parameters (i.e., changes with atmospheric depth). To model a stellar
spectrum requires a numerical representation of a stellar atmosphere, which is typically represented as ~50 layers, each with a
unique and homogeneous temperature, pressure, and chemical
composition. A blackbody spectrum is applied to the bottom
layer, and allowed to percolate upwards, with line and continuous absorption occurring in each layer, and feeding the new
spectrum into the next highest layer, until a final spectrum
emerges from the top. By iteratively comparing an observed
spectrum with this kind of synthetic spectrum, it is possible to
work backwards to determine which layers contributed to a
specific absorption line, and therefore to determine the total
number of atoms in each layer and thus in the stellar atmosphere. All stellar abundances are quoted as ratios with hydrogen, the dominant element in all stellar atmospheres.
The modeling of a stellar atmosphere is not easy. However, to
Fig. 3
Fig. 2
Variation in the iron abundance as a function of the Galactic
rotational velocity (V), demonstrating the range in metallicity for each component. Note the large scatter in the thick disk
and halo components, and especially the overlap in their
metallicities. Symbols are the same as in Figure 1.
228 C PHYSICS
IN
Sample spectra for stars in the Hyades cluster from the
ELODIE database. These stars all have the same chemical
abundances as each other and as the Sun, but due to differences in their atmospheric parameters (primarily temperature
and temperature stratification), the specific absorption lines
are clearly different.
METAL POOR STARS ... (VENN)AA
show that stellar atmosphere models yield reliable stellar
chemistries, we can compare the elemental abundances determined from the spectrum of the Sun, to those determined on
Earth from studies of meteorites. Meteorites are the left over
debris from the formation of the Sun and solar system, and are
expected to have the same birth composition as the Sun, and
therefore the same composition as the solar atmosphere. For
most elements, the agreement between the solar spectrum and
the meteorites is excellent [11]. In addition, stars with very similar properties as the Sun occur in clusters of stars (e.g., the
Hyades [12,13]; see sample spectra in Figure 3 of various stars in
this young cluster). Since clusters of stars form from the same
gas cloud at the same age, then they have the same chemistries
and thus differences in their spectra are due to differences in
their masses (thus surface properties). Analysis of the chemistry of a variety of stars in the Hyades cluster results in very
similar abundances, encouraging us that our model atmospheres analyses are able to determine elemental abundances in
stars with good accuracy.
METAL-POOR STARS
The most interesting stars to study in terms of the formation
and evolution of our Galaxy are the metal-poor stars. These
stars may have been the first stars to form in our galaxy, and
therefore are the fossilized remains of earlier epochs; or these
stars may have formed in lower mass galaxies that later merged
to form the MWG. Lower mass galaxies would simply have
less gas to convert into stars, and therefore fewer metals would
be made. The lowest metallicity stars are thought to be related
to the first stellar objects that formed in the Universe, and
therefore also have cosmological implications [14].
Today, there are dedicated efforts to find metal-poor stars, and
analyse them. Firstly, there have been several photometric and
(low resolution) spectroscopic survey which focus on certain
metal-dependent spectral features to simply find likely metalpoor stars. One example is the Hamburg-ESO survey, which
focuses on calcium features to find metal-poor stars [15]. This
survey is very sensitive to stars with metallicities less than
1/1000th that of the Sun; astronomers note this as [Fe/H] < -3,
where [Fe/H] = log(Fe/H)star - log(Fe/H)sun. This survey has
found a smooth decrease in stars with metallicity, down to
[Fe/H] = -4. Below this metallicity, there appears to be a desert
of missing objects as shown in Figure 4. Some semi-analytic
models suggest that this is the level of pre-enrichment from the
early Universe, before significant star formation began in
galaxies [16,17]. Others suggest that we just haven’t surveyed
enough stars yet and that there will be more metal-poor stars as
the searches continue.
The detailed abundances in these metal-poor stars are extremely valuable, telling us what the yields from previous star formation and supernovae explosions was like [14,19]. Metal-poor
stars must have been enriched by very low metallicity supernovae, and thankfully supernovae yields (the number of atoms
of various elements per explosion) are metallicity dependent,
e.g., low metallicity supernovae tend to produce more oxygen.
Fig. 4
The metallicity distribution of stars in the Galaxy from the
Hamburg-ESO survey. This distribution is compared with
that from four nearby dwarf spheroidal galaxies. Clearly
there is a dearth of missing objects in the Galaxy below
metallicities of [Fe/H] < -4, whereas the same occurs in the
dwarf galaxies below [Fe/H] < -3, showing that this seems to
be a common phenomenon. Figure adapted from Ref. [18].
It is also worth noting that it does not matter if the star formed
at early epochs or not; if it is metal-poor, then it is still telling
us about yields from low metallicity supernovae which can
then be applied to the first generation of supernovae. The yields
from Type II supernovae also depend on the mass of the progenitor, e.g., high mass supernovae tend to produce more oxygen and less iron. Thus, by looking at the specific ratios of the
elements, it is possible to examine the mass and metallicity
range of the supernovae that contributed to the chemistry of a
particular star.
Another important source of nucleosynthesis is during thermal
pulsing in intermediate mass stars just before they explode as
supernovae; this phase is called the Asymptotic Giant Branch
phase, and nucleosynthesis of heavy elements proceeds via the
s-process [20]. During the s-process, iron seeds capture neutrons, except the neutron capture rate is so slow that neutrons
have time to beta-decay between captures. This process is the
primary channel for forming elements like strontium and zirconium, but does not occur efficiently until stars reach metallicities near 1/100th solar (or [Fe/H] ~ -2). These yields are also
metallicity dependent.
In the Galactic halo, nearly all stars have high oxygen-to-iron
ratios, indicating they formed primarily from gas enriched only
from Type II supernovae. This is significant because there is
another type of supernova, Type Ia, which contribute irongroup elements but no alpha-elements such as oxygen. The
progenitors for this kind of supernova are the low mass stars,
which take longer to evolve and therefore there is a delay
before the interstellar medium can be enriched in iron from
Type Ia supernovae. That delay is important since significant
amounts of chemical evolution can occur in that time, including the formation of the Galactic disks.
FORMATION AND EVOLUTION OF THE
GALAXY AND THE DWARF GALAXIES
A comparison of the chemistry of stars in the distinct kinematically defined components of the Galaxy is shown in Figure 5.
For simplicity, [magnesium/iron] and [barium/yttrium] alone
are shown. These element ratios are chosen because [Mg/Fe]
traces the relative contributions of Type II supernovae (form
Mg and Fe) to Type Ia supernova (form only Fe) through the
different Galactic components. The thick and thin disks show
increasing contributions of iron without contributions to magnesium; this is expected if the stars in the disk formed later,
after a time delay when the Type Ia supernova started to pollute
the existing gas. The [barium/yttrium] ratio traces the yields
from the r- and s-process contributions. The yields of these elements are metallicity dependent (lower metallicity stars favour
barium over yttrium production), therefore this ratio is a sensitive indicator of variation in metallicity at the earliest epochs in
the evolution of a galaxy.
In the past decade, the advent of the large aperture telescopes
(8 to 10-meters) and efficient high resolution spectrographs
have made it possible to obtain high quality spectra of stars in
the nearby galaxies, as well as our own. The current technology only allows us to reach individual bright stars within
approximately 1 Mpc (3.3 million light-years) of the Galaxy;
however this encompasses most of the galaxies in the Local
Group, i.e., galaxies that are gravitationally bound to the
MWG. I have analysed spectra of bright young stars in several
of these galaxies (e.g., the Small Magellanic Cloud, M31, NGC
6822, and WLM; see references [21–24]), however it is really the
analysis of the old, metal-poor, and evolved stars that can yield
information on the early epochs of galaxy formation. This
requires analyzing the red giant stars, which are at least 1-2
magnitudes fainter than the youngest and more massive stars in
these galaxies. The brightest red giant stars can only be
analysed in detail in galaxies within ~250 kpc of the MWG.
Within this smaller range, there are dozens of extremely faint,
dwarf galaxies, which do contain many old stars. In fact, the
number of faint dwarf galaxies is steadily increasing as all-sky
surveys are examined with increasingly sophisticated noise
reducing methods [25,26].
In the hierarchical galaxy formation scenarios, these small
mass dwarf galaxies are thought to be related to the protogalactic fragments that merged to form the MWG. If so, then this can
be tested by examining the chemistry of the stars in the dwarf
galaxies to those of similar metallicity stars in the MWG. In
2003, we completed a pilot project to examine the chemistry of
a small number of individual red giant stars in four nearby
dwarf galaxies [27,28]. These four dwarf galaxies are known to
have had completely different star formation histories based on
analysis of their stellar populations. The results for those stars
are shown in Figure 6 and compared to stars in the MWG, as
well as a few additional stars from independent analyses [29-31].
The chemistry of the stars in the dwarf galaxies is not similar
to that in any distinct component of the Galaxy.
The [Mg/Fe] ratios in the dwarf galaxies tend to
be lower than stars of similar metallicity in the
Galaxy. The simplest interpretation of this is that
SN Ia contributed iron to the interstellar medium
in the dwarf galaxies when it was still at a lower
metallicity then that in the MWG. Note that this
could have occurred at the same time (or age of
the galaxy), but the lower mass of the dwarf
galaxies means that fewer stars had formed and
therefore the earlier generation of SN II did not
enrich the dwarf galaxy interstellar medium to the
same level before the SN Ia contributed. This conclusion is supported by the [Ba/Y] results where it
appears that this ratio is higher in the stars in
dwarf galaxies at a given metallicity. This occurs
when the s-process contributions are dominated
by stars of lower metallicity – thus the s-process
abundances alone show us that the stellar populaFig. 5 Element abundance ratios for stars in the Milky Way Galaxy, including magne- tions of dwarf galaxies are not chemically similar
sium (Mg), iron (Fe), barium (Ba), and yttrium (Y). Each of these elements is to those in the MWG.
sensitive to a different nucleosynthetic process, and therefore can be used to test
and constrain the chemical evolution of the Galaxy. [Mg/Fe] tests the yields
from SN II to SN Ia, while [Ba/Y] tests the contributions from s-process and
r-process neutron captures, and their metallicity dependences. Symbols are the
same as in Figure 1.
230 C PHYSICS
IN
If the Galaxy formed through hierarchical accretion of dwarf galaxies, then how is it possible that
the stars in the dwarf galaxies do not resemble the
METAL POOR STARS ... (VENN)AA
Fig. 6
Element abundance ratios for stars in seven dwarf spheroidal galaxies (black
squares), compared to stars in the Milky Way Galaxy from Figure 5. The stars
in dwarf galaxies tend to have lower [Mg/Fe] ratios at a given metallicity.
Similarly, they tend to have higher [Ba/Y] ratios at a given metallicity. These
chemical signatures suggest that no distinct component of the MWG is similar
to the stars in these dwarf galaxies.
stars in the MWG? There are several potential solutions to this
questions. The first is that we have compared the chemistries of
stars in seven different dwarf galaxies; since each of these
galaxies has had its own unique formation and evolution history, then we are mixing the signals.We really should compare
~100 stars in each dwarf galaxy to one another and to the stars
in the MWG.
This project was started in 2005 and is currently underway with
exciting preliminary results already in hand. Secondly, it is
possible that these dwarf galaxies are not representative of the
protogalactic fragments that formed the Milky Way. Possibly
those objects had higher/lower masses, higher/lower gas fractions, etc.; to answer this question requires examining stellar
chemistries in a wider variety of dwarf galaxies. Dwarf irregular galaxies have evolved in isolation and their brightest stars
can be reached with the current technologies, however their
older red giant stars cannot, and therefore this test requires
waiting for the next generation of large aperture telescopes
(e.g., the Thirty Meter Telescope). Thirdly, the stars we are
examining have formed and evolved in the dwarf galaxies
themselves; if merging happened primarily at early epochs then
perhaps only the most metal-poor stars will have
similar chemistries. Preliminary results from our
larger samples of stars in each dwarf galaxy suggest that this is possible; the chemistries of only
the most metal-poor stars are very similar
between dwarf galaxies and with both metal-poor
stars and globular clusters within our Galaxy.
However, looking back at Figure 4 in this paper
shows that the metallicity distribution of stars in
the dwarfs is not the same as that of stars in the
Galaxy. If dwarf galaxies merged only at the earliest epochs to form the MWG, as suggested by
their chemical signatures, then the question is
where did the stars with lower metallicities in the
MWG come from? Possibly dwarf “irregular”
galaxies, which are more isolated galaxies on the
outskirts of the Local Group, have lower metallicity stars than we have been finding in the nearby
dwarf spheroidal galaxies; or maybe the most
metal-poor stars are held in the lower mass, ultra
faint dwarf galaxies that have only starting to be
found from all sky surveys.
The next decade of astrophysics in the Local
Group promises many new revelations and
answers to these questions. It is an exciting time
to be doing research on stars and galaxies in our cosmic neighbourhood; in other words, think globally, act locally.
SUMMARY
The formation and evolution of the Galaxy is preserved as
chemical imprints in its stars. By examining the detailed chemical abundances of stars in the various kinematic components
of the Galaxy, it is possible to test numerical simulations and
our understanding of how the Galaxy formed. Metal-poor stars
are the real key, since their chemical imprints tend to be related to earlier times in the evolution of the Galaxy that are no
longer available to us. Some astronomers study early evolution
of galaxies by looking at galaxies at high redshifts as they are
now forming, whereas another approach is to simply crack
open the fossils that have been left behind within our own
Galaxy and its nearby neighbours. With the current and nextgeneration observational astronomy technologies, it will be
possible to examine the chemical signatures of stars throughout
the Local Group, which will be an extremely powerful tool for
deciphering the clues to cosmology that are sitting in our own
backyard.
REFERENCES:
1.
2.
3.
4.
5.
6.
7.
Eggen, O.J., Lynden-Bell, D., & Sandage, A.R., ApJ, 136, 748, (1962).
Komatsu, E., et al., submitted to ApJS (arXiv:0803.0547), (2008).
Navarro, J.F., Frenk, C.S., & White, S.D.M., ApJ, 490, 493, (1997).
Ibata, R., Gilmore G., & Irwin M., Nature, 370, 194, (1994).
McConnachie, A.W., Irwin, M.J., Ibata, R.A., Ferguson, A.M.N., Lewis, G.F., & Tanvir, N., MNRAS, 343, 1335, (2003).
Governato, F., Mayer, L., Wadsley, J., Gardner, J.P., Willman, B., Hayashi, E., Quinn, T., Stadel, J., & Lake, G., ApJ, 607, 688, (2004).
Brooks, A.M., et al., in preparation (2008).
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
19.
20.
21.
22.
23.
24.
25.
26.
27.
28.
29.
30.
31.
Venn, K.A., Irwin, M., Shetrone, M.D., Tout, C.A., Hill, V., & Tolstoy, E., AJ, 128, 1177, (2004).
Freeman, K., & Bland-Hawthorn, J. , ARAA, 40, 487, (2002).
Pritzl, B.J., Venn, K.A., & Irwin, M., AJ, 130, 2140 (2005).
Grevesse, N. & Sauval, A.J., in Solar Composition and Its Evolution, Proceedings of an ISSI Workshop, Edited by C. Fröhlich,
M.C.E. Huber, S.K. Solanki, and R. von Steiger. Published by Kluwer Academic Publishers, Dordrecht/Boston/London, p.161, (1998).
De Silva, G.M., Sneden, C., Paulson, D.B., Asplund, M., Bland-Hawthorn, J., Bessell, M.S., & Freeman, K.C., AJ, 131, 455, (2006).
Varenne, O. & Monier, R., A&A, 351, 247, (1999).
Heger, A., Woosley, S., submitted to ApJ (arXiv:0803.3161) (2008).
Beers, T.C., Christlieb, N., ARAA, 43, 531, (2005).
Scannapieco, E., Tissera, P.B., White, S.D.M., & Springel, V., MNRAS, 364, 552, (2005).
Salvadori, S., Ferrara, A., & Schneider, R., MNRAS, 386, 348, (2008).
Helmi, A., et al., ApJ, 651, 121, (2006).
Cayrel, R., et al., A&A, 416, 1117, (2004).
Herwig F., ARAA, 43, 435, (2005).
Venn, K.A., ApJ, 518, 405, (1999).
Venn, K.A., McCarthy, J.K., Lennon, D.J., Przybilla, N., Kudritzki, R.P., Lemke, M., ApJ 541, 610, (2000).
Venn, K.A., Lennon, D.J., Kaufer, A., McCarthy, J.K., Przybilla, N., Kudritzki, R.P., Lemke, M., Skillman, E.D., & Smartt, S.J., ApJ,
547, 765, (2001).
Venn K.A., Tolstoy, E., Kaufer, A., Skillman, E.D., Clarkson, S.M., Smartt, S.J., Lennon, D.J., & Kudritzki, R.P., AJ, 126, 1326,
(2003).
Belokurov V., et al., ApJ, 654, 897, (2007).
Irwin, M.J., et al., ApJ, 656L, 13, (2007).
Shetrone M.D., Venn, K.A., Tolstoy, E., Primas, F., Hill, V., & Kaufer, A., AJ, 125, 684, (2003).
Tolstoy E., Venn, K.A., Shetrone, M., Primas, F., Hill, V., Kaufer, A., & Szeifert, T., AJ, 125, 707, (2003).
Shetrone M.D., Côté, P., & Sargent, W.L.W., AJ, 548, 592, (2001).
Bonifacio, P., Sbordone, L., Marconi, G., Pasquini, L., & Hill, V., A&A, 414, 503, (2004).
Smecker-Hane, T.A., & McWilliam, A., astro-ph/0205411 (2005).
232 C PHYSICS
IN
ARTICLE DE FOND
L’ÉVOLUTION
PAR
CHIMIQUE DES GALAXIES
HUGO MARTEL
L
a composition chimique de l’univers évolue au
cours du temps. Durant les quelques minutes qui
ont suivies le big bang, les réactions nucléaires
primordiales ont convertie les protons et neutrons
en éléments légers, deutérium, hélium et lithium. Lorsque
ces réactions cessent, l’univers se compose alors d’environ 77% d’hydrogène, 23% d’hélium, avec des traces de
lithium. Les éléments plus lourds, comme le carbone, l’azote, l’oxygène ou le fer, que les astrophysiciens appellent
métaux, sont produits par les réactions nucléaires dans le
coeur des étoiles. Comme les étoiles se forment à l’intérieur des galaxies, ce sont ultimement les galaxies qui
sont responsables de l’évolution chimique de l’univers.
Durant l’évolution d’une galaxie, des étoiles se forment à
partir du milieu interstellaire. Ces étoiles évoluent, et les
plus massives d’entres elles éventuellement explosent en
supernova. Ces explosions rejètent dans le milieu interstellaire les métaux qui ont été synthétisés dans l’étoile. Le
milieu interstellaire est ainsi enrichi en métaux, et lorsque
de nouvelles étoiles se forment, elles contiendront déjà
une certaine quantité d’élements lourds. Par exemple, le
soleil contient des éléments tels que C, N, O, Ca ou Fe, qui
furent produits par des générations précédentes d’étoiles.
Ainsi, à la fois la composition chimique du milieu interstellaire et celle des étoiles évoluent au cours du temps.
Plus une étoile est vieille, plus on s’attend à ce qu’elle soit
pauvre en métaux, puisqu’elle s’est formée à une époque
où le milieu interstellaire n’avait pas encore été enrichi (on
verra que ce n’est pas toujours le cas).
Il existe deux principaux types de supernovae: les supernovae de type II (SNe II; on inclut dans cette catégorie les
SNe Ib et SNe Ic) sont produites par l’explosion d’étoiles
massives dont le temps de vie est très court (quelques millions d’années). Ces étoiles produisent principalement des
éléments α dont le numéro atomique est entre Z = 6 (carbone) et Z = 20 (calcium). Les supernovae de type Ia (SNe
Ia) sont produites par l’explosion d’étoiles de masses plus
faibles appartenant à des systèmes binaires. Le temps de
vie des progéniteurs se mesure alors en milliards d’années,
et ces supernovae produisent principalement du fer
(Z = 26). Pour quantifier les abondances des différentes
espèces chimiques, on utilise la notation suivante,
[A/B] = log10
RÉSUMÉ
L’évolution du contenu en éléments chimiques de l’univers, et la formation et l’évolution des galaxies sont des problèmes
intimement liés. Tous les éléments chimiques plus lourds que l’hélium se sont formés
à l’intérieur des galaxies. Les éléments à partir du carbone sont formés par les étoiles
contenues dans ces galaxies, alors que les
éléments plus légers (lithium, beryllium et
bore) sont formés dans le milieu diffus situé
dans les galaxies, entre les étoiles. Les éléments chimiques à leur tour influencent la
formation et l’évolution des galaxies. La
présence d’éléments lourds joue un rôle
important dans la formation d’étoiles, en
augmentant fortement le taux de refroidissement radiatif du gaz primordial. De plus la
présence d’éléments tels que le carbone,
l’azote et l’oxygène dans les étoiles affecte
le taux de réactions nucléaires dans ces
étoiles. Dans cet article, j’explique comment,
à l’aide de simulations numériques de haute
performance, on peut étudier la production
et la distribution des éléments chimiques
dans les galaxies.
⎛n ⎞
nA
− log10 ⎜ A ⎟ ,
nB
⎝ nB ⎠
(1)
où nA et nB sont les abondances des espèces A et B, et le
symbole indique les valeurs solaires. Comme le soleil
est une étoile “typique”, un rapport [A/B] positif est considéré comme élevé, et un rapport négatif est considéré
comme faible. Deux de ces rapports sont particulièrement
utiles: le rapport [Fe/H] mesure l’abondance de fer par
rapport à l’hydrogène, et peut être utilisé comme proxy
pour quantifier la métallicité. Le rapport [α/Fe] mesure
l’abondance totale de l’ensemble des éléments α par rapport au fer. Une mesure de [α/Fe] détermine l’importance
relative des différents types de supernovae, ce qui impose
des contraintes sur les modèles de formation et d’évolution de galaxies.
Hugo Martel
<[email protected].
ca>, Département de
physique, de génie
physique et d’optique, Université
Laval, Québec, QC
LES DISQUES GALACTIQUES
Il est maintenant bien établi que le disque de notre galaxie spirale, la Voie lactée, se compose en réalité de deux
disques, un disque épais et un disque mince, qui sont
imbriqués l’un dans l’autre [1,2]. Par rapport au disque
mince, les étoiles du disque épais sont cinématiquement
plus chaudes (c’est-à-dire que leur dispersion de vitesse
est plus élevée), et leur rotation traine de l’arrière par
~20 − 40 km/s [3]. Les étoiles du disque épais sont
vieilles [4-6], presque exclusivement plus agées que
L’ÉVOLUTION CHIMIQUE DES GALAXIES (MARTEL)
10 Gans 1. Malgré tout, ces étoiles sont relativement riches en
métaux, avec une distribution de metallicité qui atteint
[Fe/H] ~ −0.6 [3]. Les étoiles des disques épais et minces ont
des compositions chimiques différentes [5-13], les étoiles du
disque épais étant beaucoup plus riches en éléments α, avec un
rapport [α/Fe] systématiquement plus élevé pour une métallicité donnée. Le gradient vertical de metallicité dans le disque
épais est très faible [14,15], et de récentes observations suggèrent
que les abondances chimiques sont en fait indépendantes de la
hauteur |Z| au-dessus du plan du disque [7].
Nous savons maintenant que la plupart des galaxies spirales
possèdent un disque épais [16]. La photométrie de surface d’un
échantillon de 47 galaxies montre que ces disques épais sont
rouges [16] et donc constitués de vieilles étoiles. Plus récemment, l’étude des populations stellaires dans un échantillon de
quatre galaxies voisines a permis d’étudier les propriétés de ces
galaxies en fonction de la hauteur au-dessus du plan [17]. Toutes
ces galaxies possèdent un disque épais composé d’étoiles
rouges. Le gradient vertical de couleur est presque nul ou
légèrement positif. Le diagramme couleur-magnitude révèle
que ces disques épais consistent en une population d’étoiles
vieilles et relativement riches en métaux. Une étude du contenu
stellaire de la galaxie NGC 55 révèle que les étoiles associées
au disque épais sont vieilles (âges ~ 10 milliard d’années),
et la plupart ont une métallicité dans l’intervalle
−1.2 < [Fe/H] < −0.7 [20]. Ces résultats sont en accord avec les
conclusions de Dalcanton et Bernstein [16], et avec les observations du disque épais de la Voie lactée.
ensuite des explosions de type Ia). Lors d’une telle explosion,
l’algorithme identifie les particules de gaz voisines de la particule étoile. Ces particules de gaz sont enrichies en métaux, et
reçoivent une composante radiale de vitesse due à l’énergie
dégagée par la supernova. Lorsque ces particules de gaz formeront de nouvelles étoiles, celles-ci seront préalablement
enrichies en métaux. L’algorithme GCD+ peut ainsi suivre la
formation et l’évolution des structures dans les galaxies (telles
que bulbe, disque, halo et bras spiraux), la formation, l’évolution et la mort des étoiles, et l’évolution de la métallicité du gaz
ainsi que celle des étoiles.
Les supernovae produisent tous les élements chimiques entre
Z = 6 (carbone) et Z = 96 (curium). Cependant, il n’est pas
nécessaire d’inclure tous ces éléments dans l’algorithme. Seuls
les éléments les plus abondants sont inclus: hydrogène (H),
hélium (He), carbone (C), azote (N), oxygène (O), néon (Ne),
magnésium (Mg), silicium (Si) et fer (Fe).
Dans cette article, je présente deux exemples de simulations
numériques effectuées à l’aide de l’algorithme GCD+. Dans le
premier cas, une galaxie spirale isolée se forme par la fragmentation d’une région de densité élevée. Dans le deuxième cas,
deux galaxies spirales déjà formées entrent en collision et se
fusionnent. Ces simulations furent réalisées à l’Université
Laval, avec l’aide de mes collaborateurs Chris Brook
(University of Washington), Simon Richard (Université Laval),
Daisuke Kawata (The Observatories of the Carnegie Institution
of Washington) et Brad Gibson (University of Central
Lancashire).
L’ALGORITHME NUMÉRIQUE GCD+
Les simulations numériques présentées dans cette article furent
toutes réalisées avec l’algorithme numérique GCD+ [18]. Il
s’agit d’un algorithme Lagrangien, dans lequel le système
physique est représenté par des particules. L’algorithme utilise
trois types de particules pour représenter la matière sombre, le
gaz, et la composante stellaire 2. GCD+ combine un algorithme
à N-corps pour le calcul de la force gravitationelle avec un
algorithme Smoothed Particle Hydrodynamics pour l’hydrodynamique. La formation stellaire, l’effet rétroactif des explosions de supernovae et l’enrichissement chimique sont traités
par des règles heuristiques, qu’on appelle parfois Physique de
sous-grille. Lorsque, dans une certaine région, le gaz se retrouve dans des conditions de densité et de température favorables à la formation stellaire, une particule étoile est créée, et la
masse des particules de gaz voisines est réduite de manière à
conserver la masse et la quantité de mouvement. L’algorithme
suit l’évolution des étoiles que chaque particule étoile
représente, et cette particule subit au cours du temps une série
d’explosions de supernovae (d’abord des explosions de type II,
pour lesquelles les progéniteurs ont un court temps de vie, puis
FORMATION ISOLÉE DE GALAXIES
Une étude précédente [19] a démontré qu’une simulation
chimio-dynamique de la formation d’une galaxie avec l’algorithme GCD+ pouvait produire une galaxie spirale ayant un
disque épais et un disque mince. La présence de deux disques
distincts était indiquée par la relation entre l’âge et la dispersion de vitesse pour les étoiles autour du rayon solaire. Cette
relation montre une augmentation brusque de la dispersion de
vitesse à un âge de ~ 8Gans, parfaitement en accord avec les
observations. Dans cette simulation, le disque épais se forme
durant une période chaotique de fusions de fragments riches en
gaz à redshift 3 élevé. Ce scenario de formation du disque épais
est en accord avec les observations des disques galactiques et
extragalactiques.
Nous présentons ici quatre simulations de galaxies avec disque,
et comparons les populations stellaires de leurs disques épais
avec les observations récentes de disques épais extragalactiques [16,17,20]. Nous calculons les gradients verticaux d’âge et
de métallicité, ainsi que les couleurs. De plus, nous examinons
1. 1 Gan = 1 giga-année = 1 milliard d’années.
2. Notons que la résolution de l’algorithme est insuffisante pour résoudre les étoiles individuelles. Par conséquent, une “particule étoile” représente collectivement
un grand nombre d’étoiles.
3. Les astrophysiciens utilisent le terme redshift, ou décalage vers le rouge, pour indiquer les différentes époques cosmologiques. Le big bang correspond à un redshift z = 4, alors que le présent correspond à un redshift z = 0. Un redshift élevé correspond à une époque ancienne, alors qu’un redshift faible correspond à une
époque récente.
234 C PHYSICS
IN
L’ÉVOLUTION CHIMIQUE DES GALAXIES (MARTEL)AA
les rapports d’abondance des éléments α par rapport au fer
dans les populations stellaires du disque épais et du disque
mince, ainsi que du halo. Ceci fournira davantage de tests
importants de notre scénario de formation du disque épais, en
plus de faire le lien avec les études récentes de la formation du
halo stellaire dans les galaxies spirales.
Conditions initiales
Les conditions initiales à redshift zi consistent en une sphère
isolée composée de matière sombre et de gaz primordial
(hydrogène et hélium) dont la masse totale est Mtot, et qui
tourne lentement sur elle-même. Ceci correspond à une fluctuation de densité avec un contraste de densité δi. Comme δi > 0,
cette sphère est gravitationnellement liée, et va éventuellement
s’effondrer sur elle-même. Cet effondrement se produit à un
redshift zc donné par [21]
zc = 0.36δi (1 + zi) − 1.
Résultats
GRADIENT
VERTICAL DE LA POPULATION STELLAIRE POUR LES
ÉTOILES DU DISQUE ÉPAIS
La Figure 1 montre des images en bande I de chaque galaxie au
temps présent (z = 0), vue de face et de profil. Les quatre galaxies ont une morphologie similaire, bien que sGAL4 à première vue possède un bulbe plus important que les autres
galaxies. Chaque galaxie est dominée par un disque mince,
jeune et riche en métaux, alors que le halo stellaire est vieux et
pauvre en métaux. La densité de surface des disques suit une
loi exponentielle pour les quatre galaxies.
(2)
Des perturbations aléatoires de densité à courte échelle sont
superposées sur cette sphère. À cause de la présence de ces
perturbations, l’effondrement de la sphère ne sera pas monolithique. La sphère va d’abord se fragmenter, et les fragments
vont ensuite se fusionner lors de l’effondrement.
Nous effectuons quatre simulations (sGAL1–4) avec des
valeurs différentes de zc, Mtot et du paramètre de spin λ, qui
mesure l’importance de la rotation 4. Les valeurs des
paramètres sont données dans le Tableau 1. Les valeurs de zc
sont bien à l’intérieur des valeurs auquelles on s’attend pour la
formation de halos comparables à la Voie lactée dans un modèle cosmologique ΛCDM [22]. De plus, les perturbations aléatoires de densité incorporées aux conditions initiales sont différentes pour chaque galaxie, ce qui crée une diversité d’évolution dans nos simulations. Les perturbations sont choisies de
manière à s’assurer qu’aucune fusion majeure ne se produise à
une époque avancée (redshift z < 1). Combinées avec les
valeurs élevées de λ choisies, ces conditions initiales mènent à
la formation de galaxies avec disques dans les quatre simulations. Nous utilisons 38,911 particules de matière sombre et
38,911 particules de gaz pour chaque simulation, ce qui nous
donne une résolution comparable aux autres études récentes de
formation des galaxies spirales.
TABLEAU 1
PARAMÈTRES DES MODÈLES ET INTERVALLE D’ÂGES
POUR LA DÉFINITION DES ÉTOILES DU DISQUE ÉPAIS.
Fig. 1
Image en bande I à redshift z = 0 des quatre galaxies
simulées, vues de face (panneaux du haut) et de coté (panneaux du bas). Ces galaxies sont dominées par les étoiles
jeunes et riches en métaux du disque mince.
La présence d’un disque épais dans les quatre galaxies
simulées est révélée par la relation âge-dispersion de vitesse.
La Figure 2 montre la dispersion de vitesse dans la direction
perpendiculaire au disque (direction Z) en fonction de l’âge,
pour les étoiles situées dans le “voisinage solaire” de nos quatres galaxies. Le voisinage solaire est défini comme étant un
anneau limité par 6 kpc < RXY < 10 kpc, et |Z| < 1 kpc, où RXY
est le rayon dans le plan du disque. Le relation observée pour
les étoiles du voisinage solaire de la Voie lactée [4,23] est également indiquée par les triangles avec barres d’erreurs. La dispersion de vitesse observée est relativement constante pour les
dernières ~ 9 Gans 5, mais montre une augmentation brusque à
une époque de ~ 10 Gans dans le passé. Les étoiles plus vieilles
avec une dispersion de vitesse élevée sont identifiées comme
appartenant au disque épais. Les dispersions de vitesse des quatres galaxies ont qualitativement le même comportement, avec
des plateaux interrompus par des augmentations brusques à des
périodes entre 8 et 10 Gans dans le passé. Ceci implique que
chaque galaxie possède un disque épais. Cette augmentation
brusque est plus récente pour sGAL1 et sGAL2 (~ 8 Gans dans
le passé) que pour sGAL3 et sGAL4 (~ 10 Gans). Cette différence est principalement due aux différences entre les redshifts d’effondrement zc dans les conditions initiales.
4. Le paramètre de spin est le rapport entre l’énergie cinétique de rotation et l’énergie de liaison.
5. Une étude plus récente [24] montre une augmentation de la dispersion de vitesse avec l’âge pour les étoiles du disque mince, mais cela est sans conséquence pour
les résultats présentés ici.
matiquement par la perturbation des fragments, et acquiert une
dispersion de vitesse élevée. Par conséquent, les étoiles formées dans le disque durant cette période auront une dispersion
de vitesse élevée, et seront ultimement identifiées comme étant
des étoiles du disque épais.
Fig. 2
Dispersion de vitesse des étoiles dans la direction perpendiculaire au plan (direction Z) en fonction de leur âge, pour les
quatre galaxies simulées: sGAL1 (+), sGAL2 (∗), sGAL3 (o)
et sGAL4 (H). Le calcul inclut les étoiles situées dans la
région définie par 6 kpc < RXY < 10 kpc et |Z| < 1 kpc. Les triangles avec barres d’erreurs montrent les résultats obtenus à
partir d’observations d’étoiles situées dans le voisinage du
soleil. On remarque dans tous les cas une augmentation
brusque de la dispersion de vitesse à un âge > 8 Gans. Cette
augmentation brusque est la signature de la formation du
disque épais.
Nous utilisons ces accroissements de la dispersion de vitesse
comme indicateurs de l’époque de formation du disque épais.
La Figure 3 montre quatre stages différents de l’évolution de
sGAL1 durant l’époque de formation du disque épais. Cette
époque est caractérisée par de multiples fusions de fragments
riches en gaz. Nous confirmons que le rapport entre la masse
de gaz et la masse stellaire dans ces fragments est élevé à cette
époque [19]. Le gaz, qui contient une quantité importante de
moment cinétique, se retouve dans une structure de la forme
d’un disque. Cependant, le disque gazeux est réchauffé ciné-
Cette époque est également caractérisée par une formation stellaire rapide. Le taux global de formation stellaire (SFR) en
fonction du temps est indiqué sur la Figure 4. L’époque de formation du disque épais montrée sur la Figure 3 correspond
assez bien au maximum du SFR. Ce maximum est atteint plus
tard pour sGAL1 et sGAL2 que pour sGAL3 et sGAL4, ce qui
implique que les disques épais de sGAL3 et sGAL4 se sont formés plus tôt que ceux de sGAL1 et sGAL2, en accord avec la
Figure 2. Ce n’est qu’après la formation du disque épais que le
disque mince commence à se former [19].
Fig. 4
Taux de formation stellaire (SFR) en fonction de l’âge pour
les quatre galaxies simulées. Le maximum de la formation
stellaire correspond à une époque de fusions entre fragments
riches en gaz, que nous associons à la formation du disque
épais.
L’intervalle d’âge pour la formation du disque épais est indiqué
dans le Tableau 1. Nous avons choisi cet intervalle d’âge en
nous basant sur la dispersion de vitesse (Figure 2), les images
(Figure 3) et le taux de formation stellaire (Figure 4). Nous
utilisons un critère supplémentaire pour distinguer les étoiles
du disque épais de celles du halo, qui se forment durant la
même période. Les étoiles doivent avoir une vitesse de rotation
supérieure à 50 km/s pour appartenir au disque épais.
Fig. 3
Densité d’étoiles (panneaux du haut) et de gaz (panneaux du
bas) pour sGAL1 durant l’époque de la formation du disque
épais, vue de face. On montre quatre temps différents, correspondants à des âges de 9.8 − 8.5 Gans. Cette époque correspond au temps où la relation dispersion de vitesse-âge montre une croissance subite, indiquant la présence du disque
épais (Figure 2), et est caractérisée par des fusions de fragments riches en gaz. Au début de cette séquence, plusieurs
fragments sont présents, alors qu’à la fin une galaxie s’est
formée.
236 C PHYSICS
IN
Les étoiles du disque épais identifiées dans nos simulations ont
des propriétés similaires à celles observées dans le disque épais
de la Voie lactée. La rotation du disque épais traine de l’arrière
par rapport aux étoiles du disque mince par 20 − 30 km/s. Nous
obtenons des échelles de hauteur de 1.3, 1.4, 1.1, et 1.4 kpc
pour les disques épais de sGAL1–4, comparé à des échelles de
hauteur de 0.52, 0.55, 0.6 et 0.45 kpc pour les disques minces.
Dans ce qui suit, nous allons démontrer que les étoiles du
disque épais ont également des métallicités caractéristiques
observées dans la Voie lactée et les autres galaxies spirales.
La question clé dans les simulations de disques galactiques est
la variation de la métallicité, de l’âge, et ainsi des couleurs des
étoiles du disque épais en fonction de la hauteur |Z| au-dessus
du plan du disque, ainsi que les différences entre les rapports
d’abondance des étoiles du disque épais et ceux des étoiles du
disque mince que se forment plus tard. Dans les Figures 5 à 8,
nous analysons les étoiles du disque épais dans la région
6 kpc < RXY < 10 kpc. La Figure 5 montre la métallicité des
étoiles en fonction de |Z|. Les étoiles de sGAL1, sGAL2, et
sGAL4 ont des métallicités dans l’intervalle de +[Fe/H], entre
−0.5 et −0.6, ressemblant de près au disque épais de la Voie lactée, alors que les étoiles de sGAL3 ont une métallicité nettement plus faible, +[Fe/H], ~ −0.8. Ceci pourrait être du au fait
que le disque épais de sGAL3 et moins massif que celui des
autre galaxies. Les étoiles du disque épais n’ont aucun gradient
vertical de métallicité.
Fig. 5
Fig. 6
Moyenne du logarithme de l’âge (Gans) en fonction de la
hauteur |Z| au-dessus du disque, pour les étoiles du disque
épais. On utilise les mêmes symboles que sur la Figure 2. Les
étoiles du disque épais dans chaque galaxie sont vieilles, avec
peu de variation avec la hauteur.
Fig. 7
Moyenne de l’indice de couleur V − I en fonction de la hauteur |Z| au-dessus du disque, pour les étoiles de la branche des
géantes (RGB/AGB) du disque épais. On utilise les mêmes
symboles que sur la Figure 2.
Métallicité moyenne +[Fe/H], en fonction de la hauteur |Z|
au-dessus du disque, pour les étoiles du disque épais. On
utilise les mêmes symboles que sur la Figure 2. Les étoiles du
disque épais de sGAL1, sGAL2 et sGAL3 ont une métallicité entre −0.5 et −0.6, alors que celles de sGAL4 ont une
métallicité plus faible, +[Fe/H], ~ −0.8. Il n’y a que très peu
de gradient, pour les quatre galaxies.
La Figure 6 montre l’âge des étoiles du disque épais en fonction de |Z|. Les variations de l’âge avec la hauteur sont très
faibles. Il peut sembler bizarre que les étoiles du disque épais
de sGAL4 soient en moyenne plus vieilles que celles de
sGAL3, apparemment en contradiction avec les époques de
formation données dans le Tableau 1. Ceci est causé par la
baisse rapide du taux de formation stellaire dans sGAL4, tel
que l’indique la Figure 4, qui implique qu’une plus grande
fraction des étoiles du disque épais de sGAL4 étaient déjà formées à l’intérieur des fragments, avant que ceux-ci se fusionnent pour former le disque épais.
Comme les étoiles du disque épais n’ont pas de gradient vertical d’âge et de métallicité, elles n’ont pas non plus de gradient
vertical de couleur. La Figure 7 montre la valeur moyenne de
l’indice de couleur V − I en fonction de |Z|. On ne distingue
aucun gradient vertical de couleur. Ces résultats sont en accord
avec les couleurs et l’absence de gradients de couleur
observées dans les galaxies spirales [17].
La Figure 8 montre les indices de couleurs B − R et R − K des
étoiles du disque épais. Ces couleurs sont obtenues en intégrant
la luminosité de toutes les étoiles du disque épais qui sont
encore présentes à la fin de la simulation. Ces couleurs sont relativement constantes avec la hauteur, avec B − R ~ 1.4 − 1.5 et
R − K ~ 1.9 − 2.2.
Ces couleurs nous permettent une comparaison directe avec les
Figures 3 et 6 de Dalcanton et Bernstein [16]. Loin du plan,
Dalcanton et Bernstein [16] trouve un intervalle de couleur relativement faible, avec B − R ~ 10 − 1.4 et R − K ~ 2.0 − 2.6. Les
galaxies que nous simulons sont plus massives que les
47 galaxies étudiées par Dalcanton et Bernstein [16]. Malgré
tout, les disques épais de nos galaxies ont des couleurs qui peu-
Fig. 9
Fig. 8
Moyenne des indices de couleur B − R et R − K en fonction
de la hauteur |Z| au-dessus du disque, pour les étoiles du
disque épais. On utilise les mêmes symboles que sur la
Figure 2.
vent expliquer les populations stellaires enveloppant les galaxies observées par Dalcanton et Bernstein [16], qui sont interprétées comme appartenant à un disque épais.
[α/FE] VS. [FE/H] POUR LES ÉTOILES DU DISQUE MINCE, DU DISQUE
ÉPAIS ET DU HALO.
Nous comparons maintenant les rapports d’abondance des éléments α par rapport au fer pour les differentes composantes des
galaxies, soit le disque mince, le disque épais et le halo. Afin
de pouvoir comparer les résultats de nos modèles avec les
observations, nous examinons les étoiles situées à l’intérieur du
voisinage solaire. Les étoiles du disque mince sont définies
comme étant celles qui sont plus jeunes que 7 milliards d’années et qui tournent plus vite que 50 km/s. Les étoiles du halo
sont définies comme étant les étoiles rétrogrades formées avant
la fin de l’époque de formation du disque épais, tel que définie
dans le Tableau 1. Notons que les étoiles déjà formées dans les
fragments qui plus tard se fusionnent durant la formation du
disque épais ont une forte tendance à se retrouver dans le
halo [25] 6. La Figure 9 montre l’abondance en éléments α des
trois composantes pour la galaxie sGAL1 (les résultats pout les
trois autres galaxies sont très similaires). L’échelle de l’axe
[Fe/H] a été choisie pour inclure environ 90% des étoiles de
chaque composante.
Dans toutes nos simulations, les étoiles du disque épais ont une
abondance en éléments α plus grande que celles du disque
mince, même lorsqu’elles ont la même métallicité [Fe/H]. Ces
abondances du disque épais sont caractéristiques de l’enrichissement dominant par les SNe II. Ces résultats cadrent
bien avec les observations récentes [5-13]. Nous nous devons de
mentionner certaines incertitudes dans nos modèles, concer-
[α/Fe] versus [Fe/H] pour les étoiles du “voisinage solaire”
de sGAL1. Les différent symboles représentent les étoiles du
disque mince (∗), du disque épais (~) et du halo (‘).
nant la formation stellaire, le taux de production des éléments
par nucléosynthèse, et particulièrement les échelles de temps
des SNe Ia, qui sont pas très bien expliquées par les modèles
théoriques. Néanmoins, il est très encourageant de constater
que les différences entre les schémas d’abondances du disque
mince et du disque épais de la Voie lactée peuvent être reproduits avec notre scénario de formation du disque épais.
Les étoiles du halo ont également des abondances élevées
d’éléments α, comme celles du disque épais. Les étoiles formées dans les fragments riches en gaz vont préférentiellement
s’accréter au halo et non au disque [25]. Celles formées dans les
fragments accrétés à redshift élevé auront été enrichies principalement par les SNe II. Les étoiles du halo ont tendance à
avoir des valeurs de [α/Fe] légèrement plus élevées que celles
du disque épais. La raison semble être que les SNe II produisent des rapports [α/Fe] plus élevés [27]. Les étoiles du
disque mince, qui se forment plus tard durant la période calme
qui suit la formation du disque épais, sont formées à partir de
matériel enrichi davantage par les SNe Ia, qui produisent du
fer. Elles ont donc un rapport [α/Fe] plus faible.
Discussion
Plusieurs scénarios ont été proposés pour expliquer l’origine du
disque épais de la Voie lactée [19,28,29], qui est maintenant établi
comme étant une composante séparée du disque mince. Ces
différents scénarios suscitent un intéret particulier depuis que
de récentes observations suggèrent que les disques épais sont
omniprésent dans les galaxies spirales, et que leurs âges
avancés contient peut-être la clé qui expliquera la formation de
telles galaxies. Un scenario a été proposé [19], dans lequel la
majorité des étoiles du disque épais se forment à redshift élevé,
durant une période de fusions multiples de fragments riches en
gaz précédant la formation du disque mince. Le but des quatre
simulations présentées dans cette article était de tester ce scénario.
Des observations d’étoiles individuelles dans quatres galaxies
vues de côté ont confirmé que (a) les disques épais apparaissent
6. La tendance des étoiles accrétées à être circularisées et ainsi faire partie des composantes du disque, tel que révélée par Meza et al. [26], se produit seulement
après que le disque se soit déjà formé. On s’attent à ce que ces étoiles soient rares dans la Voie lactée, parce que de telles étoiles devraient avoir une faible
métallicité, et la fonction de métallicité des étoiles du voisinage solaire suggère peu d’étoiles ayant une faible métallicité.
238 C PHYSICS
IN
fréquemment dans ces galaxies, (b) les étoiles du disque épais
sont vieilles et relativement riches en métaux, et (c) il n’y a pas
de gradient vertical de metallicité important [17]. Ces observations ne supportent pas la formation de disques épais par l’accrétion d’étoiles provenant de satellites (puisque ces étoiles
seraient pauvres en métaux), ou par le réchauffement lent du
disque mince (ce qui produirait un gradient vertical de couleur,
et de plus certaines étoiles de disque épais seraient jeunes).
L’absence de détection d’un gradient vertical de métallicité
place également de sérieuses contraintes sur un scénario basé
sur un effondrement dissipatif lent. Nous avons montré dans les
Figures 5 et 6 que notre scénario de formation du disque épais
est en accord avec les observations. En particulier, il n’y a pas
de variation de la métallicité ou de l’âge, et donc de la couleur,
avec la hauteur. C’est parce que, dans notre scénario, le disque
épais se forme à partir de gaz déjà bien enrichi en métaux,
durant une période de temps relativement courte.
L’enrichissement durant et après les fusions, ainsi que l’accumulation continue de gaz [30] élimine les différences de métallicité dans les fragments qui forment le disque épais, ce qui
produit une population stellaire vieille et riche en métaux relativement homogène, indépendante de la hauteur.
Une étude de 6 galaxies spirales vues de côté avec une population stellaire résolue supporte également l’existence d’une population appartenant à un disque épais ou un halo [31], mais cette
analyse préliminaire des couleurs des étoiles géantes de la
branche rouge indique que cette enveloppe stellaire est plus
pauvre en métaux que celles trouvées par Mould [17] et
Davidge [20], avec des métallicités −1.2 < [Fe/H] < −1.7. Ceci
suggère une diversité de la population stellaire des disques
épais parmi les différentes galaxies spirales. Nos quatre galaxies simulées, qui sont massives, n’ont pas une aussi faible
métallicité. Un plus grand échantillon de galaxies simulées
avec des masses différentes pourraient expliquer de telles
observations.
masse relativement faible (souvenons-nous de la relation
masse-métallicité bien établie pour les galaxies), et la courte
période de temps disponible pour la formation stellaire avant
les fusions, assure que ces étoiles sont pauvres en métaux. Ceci
peut expliquer la présence des étoiles les plus vieilles et plus
pauvres en métaux du halo. La Figure 9 montre que les étoiles
du halo sont également riches en éléments α. Ce scénario de
formation des étoiles du halo est supporté par une étude semianalytique récente [32] qui montre que les fragments qui se
fusionnent à redshift élevé sont probablement similaires à des
galaxies naines irrégulières relativement massives
(~ 5 H 1010M ) riches en éléments α.
Mentionnons que l’accrétion d’étoiles après que le disque soit
formé [33] peut aussi jouer un rôle en contribuant à la population du disque épais [26,34-37]. Nous savons déjà qu’une telle
accrétion joue un rôle dans la formation du halo stellaire [34,38,39].
Une étude récente a révélé la présence d’un disque en contrerotation dans NGC 227 [40]. L’existence de disques en contrerotation invaliderait les modèles dans lesquels les disques épais
se forment purement par effondrement monolithique ou par le
réchauffement d’un disque mince, et favoriserait fortement les
modèles d’accrétion ou de fusion. Les conditions initiales de
nos modèles, dans lesquelles une rotation de corps solide est
impartie à notre sphère initiale, impliquent que nous ne pouvons pas tester directement l’existence de disques en contrerotation. Dans nos simulations, tous les fragments massifs qui
se fusionnent durant l’époque de la formation du disque épais
ont une rotation prograde. Cependant, il ne serait pas surprenant que durant une période de fusions multiples, certains
fragments puissent être en contre-rotation. Donc, notre scénario de formation du disque épais demeure un bon candidat
pour expliquer de telles observations.
COLLISIONS ENTRE GALAXIES SPIRALES
Dans notre scénario de formation du disque épais, les fragments qui se fusionnent à redshift élevé ont été enrichis principalement par les SNe II. Nous avons montré dans la Figure 9
que notre scénario explique naturellement l’abondance élevée
d’éléments α observée dans le disque épais comparée au disque
mince [7,8,11] . Ceci est du au fait que, dans notre scénario, les
étoiles du disque épais se forment plus tôt que celles du disque
mince, qui auront la chance d’être enrichies davantage par les
SNe Ia. L’autre facteur important affectant l’abondance en éléments α est le taux élevé de formation stellaire durant l’époque
de fusion durant laquelle le disque épais se forme. Nous savons
que le disque épais a une histoire de formation stellaire plus
intense que le disque mince [7,9].
Notre scénario de formation du disque épais cadre bien avec la
formation d’un halo stellaire agé, pauvre en métaux et riche en
éléments α. Dans Ref. [25], l’accrétion d’étoiles par le halo
dans les fusions riches en gaz avait été démontrée comme étant
nécessaire pour former des halos de faibles masses et faibles
métallicités. Donc, beaucoup de ces étoiles du halo se forment
dans les fragments, avant l’époque de fusions multiples. La
Une collision majeure entre deux galaxies spirales, suivie de la
fusion de ces galaxies, peut parfois résulter en la formation
d’une nouvelle galaxie spirale [41]. Ceci se produit pour un
large éventail de rapports de masses, d’orbites et de vitesses de
rotation des galaxies progénitrices [42]. La découverte que les
disques épais sont fréquents, et possiblement omniprésents
dans les galaxies spirales [17,43], et leur âges avancés, implique
que ces disques contiennent des indices importants sur la formation de telles galaxies. Elle implique également que les
galaxies spirales formées par une fusion majeure devraient
toutes posséder un disque épais. Les galaxies impliquées dans
ces fusions étaient présumément plus riches en gaz à des époques plus anciennes, puisque le gaz est converti en étoiles au
cours du temps. La possibilité que des fusions riches en gaz
puissent jouer un rôle essentiel dans la formation des galaxies
spirales a récemment gagné en popularité [19,44]. Que les
fusions à redshift élevé soient riches en gaz est aussi en accord
avec d’autres contraintes, incluant un halo de faible masse pauvre en métaux [25], les schémas d’abondances chimiques des
étoiles du halo [45,46], et le moment cinétique des disques [47].
Le rapport entre les abondances des éléments α et celle du fer
peut être utilisé comme une horloge pour identifier les différentes échelles de temps de formation pour les différents
types de galaxies, ainsi que pour les différentes composantes de
la Voie lactée. Les supernovae SNe II, qui suivent de près la
formation stellaire à cause du cours temps de vie de leurs
progéniteurs, produisent une grande quantité d’éléments α. Le
fer, quant à lui, est produit principalement par les supernovae
SNe Ia, pour lesquelles les explosions sont retardées. Donc, le
rapport [α/Fe] élevé qui caractérise les galaxies elliptiques
implique une formation rapide de ces galaxies, alors que le rapport [α/Fe] faible que l’on trouve dans les galaxies spirales et
les galaxies naines implique une formation plus étendue dans
le temps. Dans la Voie lactée, le halo stellaire de faible métallicité ([Fe/H] ~ −1.5) a un rapport [α/Fe] ~ 0.4, alors que le
disque épais, qui a une métallicité relativement élevée
([Fe/H] ~ −0.6) a également une abondance élévée d’éléments
α par rapport aux abondances solaires. Récemment, plusieurs
études [7,48] ont révélé que les composantes chaude et froide du
disque de la Voie lactée n’ont pas le même rapport [α/Fe], fournissant ainsi un indice sur la formation de ces composantes.
Nous avons suggéré que le taux élevé de formation stellaire
dans les fusions de galaxies riches en gaz produit une métallicité élevée avec une abondance élevée d’éléments α [44]. De
tels événements pourrait jouer un rôle central dans la formation
du disque épais de la Voie lactée. Ici, nous examinons en détail
un cas particulier, et suivons l’évolution du rapport [α/Fe]
durant l’époque de la fusion.
Détails de la simulation et résultats
Nous avons utilisé l’algorithme GCD+ pour réaliser une simulation chimio-dynamique de la fusion de deux galaxies spirales
riches en gaz. Les conditions initiales sont illustrées sur la
Figure 10. Elles consistent en deux galaxies ayant chacune un
disque exponentiel de gaz situé à l’intérieur d’un halo de
matière sombre. Ces galaxies sont créées avec le logiciel
GalactICS [49], et sont essentiellement stables, dans le sens que
leurs profils de densité, potentiels et ellipsoïdes de vitesses ne
changent pas de manière significative lorsque ces galaxies
évoluent en isolation. La plus grande galaxie a une masse de
5 H 1011M . Le rapport des masses est de 2:1, les longueurs
d’échelles des disques sont de 4.5 kpc et 3.1 kpc et chaque
galaxie a une fraction baryonique de 17%. Chaque galaxie est
constituée de 40,000 particules de gaz et 100,000 particules de
matière sombre. La petite galaxie s’approche avec son axe de
rotation incliné de 17o par rapport à l’axe de rotation de la
grande galaxie. La rotation des deux galaxies est prograde par
rapport à l’orbite du système. L’énergie cinétique orbitale est
de 1.7 H 1044 ergs. Le système est lié gravitationellement, et
son paramètre de spin λ est égal à 0.04. Les disques galactiques
évoluent et forment des étoiles avant la collision, et au moment
de la collision 91% de la matière baryonique se trouve sous
forme de gaz. Initiallement, le gaz a une métallicité de
[Fe/H] = −4 et une abondance en éléments α de [α/Fe] = 0.35.
Nous suivons l’évolution du système pour 1.5 milliard d’années.
240 C PHYSICS
IN
Fig. 10 Géométrie des conditions initiales. L’axe des Y pointe vers
l’arrière. Gal1 est initialement au repos, alors que Gal2 est
initialement en mouvement dans la direction +Y. Les lignes
pointillées indiquent les axes de rotation des galaxies. Gal1
est située dans le plan X − Y, avec son axe de rotation dans la
direction Z. L’axe de rotation de Gal2 est dans le plan X − Z,
à un angle θ = 17o par rapport à l’axe des Z. Les deux galaxies tournent dans le sens horaire lorsque vues par le dessus,
par conséquent les bords de gauche s’éloignent alors que
ceux de droite se rapprochent. Les bords des disques sont
situés à deux longueurs d’échelle.
La fusion riche en gaz produit une galaxie finale qui possède
un disque. La Figure 11 montre une carte de luminosité en
bande B de la galaxie ainsi formée après 1.5 Gans, vue de face
(panneaux du haut) et de profil (panneaux du bas). Les panneaux de droite montrent toutes les étoiles. Les panneaux du
centre montrent les étoiles qui sont formées avant ou durant la
fusion, que nous appellerons par la suite étoiles fusion. Les
panneaux de gauche montrent les étoiles qui sont formées après
la fusion, que nous appellerons par la suite étoiles disque. La
vue de face montre clairement que cette simulation a produit
une galaxie à anneau, ce qui indique que les fusions riches en
gaz de disques progrades peuvent produire de telles galaxies.
Ceci est particulièrement intéressant si on considère que la
fréquence des galaxies à anneau augmente rapidement avec le
redshift [50].
Fig. 11 Luminosité en bande B de la distribution finale, vue de face
(panneaux du haut) et de côté (panneaux du bas), pour toutes
les étoiles (panneaux de droite), les étoiles fusion (formées
avant ou pendant la fusion; panneaux de milieu), et les étoiles
disque (formées après la fusion, panneaux de gauche).
La Figure 12 montre les profils de luminosité en bande B. À
partir de ces courbes, on détermine des longueurs d’échelle de
5.1 kpc pour les étoiles fusion et 4.1 kpc pour les étoiles
disque. Un bulbe central est visible à l’intérieur des ~ 3 kpc
centraux, en accord avec les résultats de Robertson et al. [42].
Notons que nos conditions initiales n’incluent pas une composante sphéroïdale. Toute étoile située dans une telle composante se retrouverait dans la composante sphéroïdale de la
galaxie finale, et donc nos conditions initiales idéalisées sont
partiellement responsables de l’absence de composante
sphéroïdale importante dans la galaxie finale. Nos résultats
demeurent valides tant que toute composante sphéroïdale initiale a une masse trop petite pour affecter la dynamique de la
fusion.
Fig. 13 Taux de formation stellaire en fonction de l’ âge. Le sursaut
de formation d’étoiles qui se produit durant la fusion atteint
un maximum de 380M /an. La ligne pointillée indique le
début de la fusion. La ligne en tirets indique la division entre
la formation des étoiles fusion et celle des étoiles disque.
Fig. 12 Profils de luminosité en bande B pour les étoiles fusion (H) et
les étoiles disque (+). Les lignes pointillées sont des ajustements entre 2.5 et 15 kpc, et indiquent des longueurs
d’échelle de 5.1 et 4.1 kpc (en ignorant une région dense
associée à un anneau dans les étoiles disque).
Le taux de formation stellaire (Figure 13) montre un sursaut
avec un maximum de 380 M /an, durant la fusion. La fin de ce
sursaut est utilisée pour diviser les étoiles en étoiles fusion et
étoiles disque. Avant la fusion, le taux de formation stellaire est
de l’ordre de 30 M /an. Après la fusion, le taux de formation
stellaire tombe en dessous de 10 M /an après un milliard d’années. La masse totale des étoiles formées avant, durant et après
la collision est de 6.3 H 109M , 33 H 109M et 19 H 109M .
Dans la région du disque, 4 kpc < RXY < 10 kpc et |Z| < 1 kpc,
les étoiles fusion et étoiles disque ont une masse totale de
3.5 H 109M et 4.4 H 109M . Les étoiles fusion forment une
structure plus épaisse que les étoiles disque. Ces deux composantes se distinguent également lorsqu’on considère leurs
fonctions de métallicité, comme le montre la Figure 14. Les
étoiles fusion ont une métallicité maximale de l’ordre de
[Fe/H] ~ −0.8, alors que les étoiles disque ont une métallicité
maximale de [Fe/H] ~ −0.2 . La longue queue à faibles métallicités pour les étoiles fusions est due à la différence entre les
étoiles formées avant et et celles formées durant le sursaut de
formation d’étoiles.
Le panneau du haut de la Figure 15 montre la vitesse de rotation des étoiles fusion et étoiles disque. Sans surprise, les
étoiles fusion ont une vitesse de rotation nettement plus faible.
Fig. 14 Distribution de la métallicité pour les étoiles situées dans la
région 4 kpc < RXY < 10 kpc et |Z| < 1 kpc dans la galaxie
finale. Ligne pointillée: étoiles disque; ligne à tirets: étoiles
fusion; ligne solide: toutes les étoiles.
Il est important de mentionner qu’une partie des étoiles fusion
sont en contre-rotation. Nous avons mesuré de nouveau la rotation, en excluant les étoiles en contre-rotation, et avons trouvé
que leur rotation demeure beaucoup plus faible que celle des
étoiles disque. Notre définition de l’époque de transition entre
les étoiles fusion et les étoiles disque est en accord avec le
changement de la dispersion de vitesse, comme on le voit dans
le panneau du bas de la Figure 15, qui montre la dispersion de
la vitesse de rotation en fonction du temps de formation des
étoiles.
La Figure 16 montre les abondances chimiques, pour une
tranche dans le plan de la galaxie, à |z| < 1 kpc. Le rapport
[α/Fe] en fonction du rayon (en haut à gauche) montre que les
étoiles fusion (H) ont une valeur de [α/Fe] ~ 0.35 indépendamment du rayon (bien que dans les régions intérieures associées
au bulbe, la valeur soit un peu plus faible), alors que les étoiles
disque (+) ont une valeur ~ 0.15 dex plus faible. Le graphique
de [α/Fe] en fonction de [Fe/H] (en haut au centre) est intéressant. Les étoiles fusion maintiennent un rapport [α/Fe] plus
élevé, même quand leur rapport [Fe/H] a atteint les valeurs
disque montrent une légère diminution de [Fe/H] avec la hauteur.
Fig. 16 Dans tous les panneaux les symboles H désignent les étoiles
fusion et les symboles + désignent les étoiles disque.
Panneaux du haut, [α/Fe] en fonction du rayon (gauche), du
rapport [Fe/H] (centre), et de la hauteur au-dessus du plan du
disque (pour les étoiles avec 4 kpc < RXY < 10 kpc, droite).
En bas à gauche, [α/Fe] en fonction du temps. En bas au centre, [Fe/H] en fonction du temps. En bas à droite, [Fe/H] en
fonction de la hauteur au-dessus du plan du disque (pour les
étoiles avec 4 kpc < RXY < 10 kpc ).
Discussion
Fig. 15 Panneau du haut: Courbes de rotation pour les étoiles fusion
(H) et les étoiles disque (+). Les étoiles en contre-rotation ne
sont pas incluses dans ces courbes, et leur courbe de rotation
est calculée séparément (∗). Panneau du bas: Dispersion de
vitesse dans la direction de rotation en fonction de l’ époque
de formation des étoiles. La ligne en tirets correspond à la fin
de la fusion, tel qu’indiquée sur la Figure 13.
solaires, alors que les étoiles disque qui sont formées avec des
valeurs de [Fe/H] ~ −0.5 ont des valeurs relativement faibles
de [α/Fe]. L’examen du panneau en bas à gauche, où nous
montrons l’évolution de [α/Fe] au cours du temps, aide à expliquer ce résultat. Les premières étoiles ont un rapport
[α/Fe] ~ 0.35, proche des conditions initiales, qui diminue au
cours du temps avant la fusion. La valeur de [α/Fe] augmente
durant la période de sursaut de formation stellaire, ce qui est
indiqué par le saut entre le deuxième et le troisième point, pour
ensuite diminuer jusqu’à la valeur trouvée dans le disque.
Durant le sursaut de formation stellaire, un grand nombre de
SNe II produisent une réserve d’éléments α qui permet de
maintenir un rapport [α/Fe] élevé, même si le contenu en fer du
réservoir de gaz augmente (panneau en bas au centre). Après la
fusion, la pollution par les SNe Ia devient importante, et le rapport [α/Fe] est maintenu à une valeur constante. Le panneau en
haut à droite montre que ni les étoiles fusion ni les étoiles
disque ont un gradient vertical de [α/Fe], alors que les étoiles
242 C PHYSICS
IN
Deux disques se forment naturellement lors d’une fusion riche
en gaz, un disque épais et un disque mince. Le disque épais
consiste en étoiles formées avant et durant la collision, bien que
certaines étoiles formées avant la collision se retrouvent dans
le halo stellaire comme dans les simulations de Springel et
Hernquist [41]. Le disque mince se forme rapidement à la fin de
la fusion. Le taux élevé de formation stellaire déclenchée par
de telles fusions aide à expliquer les abondances chimiques des
étoiles. Le sursaut de formation stellaire produit une augmentation rapide de la métallicité (Figure 16, panneau en bas au
centre), et la courte échelle de temps de formation stellaire
assure que les étoiles formées sont enrichies principalement par
les SNe II. Par conséquent les “étoiles fusion” maintiennent un
rapport [α/Fe] élevé, même après que leur rapport [Fe/H] ait
atteint des valeurs solaires. Les étoiles formées après la fusion
(les “étoiles disque”) se forment avec des valeurs faibles de
[α/Fe], même celles formées avec un faible [Fe/H] ~ −0.5. Ces
étoiles se forment dans un disque mince durant la période
calme qui suit la fusion, et ont une faible dispersion de vitesse
(Figure 15, panneau du bas).
Une simulation précédente a montré que l’échelle de longueur
du disque épais était plus courte que celle du disque mince [19].
Ici, nous obtenons le résultat inverse. La population chaude de
la fusion a un profil exponentiel avec une longueur d’échelle
plus grande que celle de la population du disque qui se forme
plus tard. Ce résultat est en accord avec les observations, qui
trouvent que l’échelle de longueur du disque épais est systématiquement plus grande que celle du disque mince [40]. Ceci peut
favoriser une fusion riche en gaz comme étant une phase
importante de la formation des galaxies spirales. Des simulations futures vont déterminer si les disques vieux et chauds,
avec une longueur d’échelle plus grande que celle des disques
jeunes et froid, sont produits pour un grand éventail de
paramètres de fusion. Il faut interpréter les résultats prudemment, puisque la croissance subséquente du disque mince par
déposition de gaz (absent de nos simulations) pourrait augmenter l’échelle de longueur de disque mince [51]. Yoachim et
Dalcanton [40] trouvent que les galaxies spirales de faible masse
ont des rapports de masse (disque épais: disque mince) plus
grand. Ils y voient une évidence de la formation du disque épais
par accrétion directe d’étoiles, puisque les progéniteurs de
galaxies de faible masse vont plus facilement éjecter leur gaz
hors de leur faible puit de potentiel avant la fusion. Malgré
tout, les observations suggèrent que les galaxies de faible
masse sont en fait plus riches en gaz, autant à redshift faible [52]
qu’à redshift élevé [53], favorisant les fusions riches en gaz
comme étant responsables des rapports de masse disque
épais:mince élevés dans les galaxies de faible masse. De plus,
la croissance du disque mince dans les petites galaxies est peutêtre contrôlée par leurs faibles densités [54] et l’effet rétroactif
des supernovae. Le rapport de masse élevé du disque épais au
disque mince dans les galaxies de faible masse résulterait alors
d’une croissance plus faible du disque mince.
Notre étude idéalisée n’inclut pas l’accrétion de gaz à partir du
milieu intergalactique, qui joue un rôle important dans la formation des disques galactiques. De plus, dans le scénario
hiérarchique de formation de structures, il est probable que
plusieurs fusions riches en gaz se soient produites durant la for-
mation d’une galaxie spirale [44]. Un mode d’accrétion froide à
partir de structures filamentaires se produit également dans ce
scénario. Cependant, notre étude supporte l’idée que l’accrétion violente de galaxies riches en gaz joue un rôle central dans
la formation du disque épais. En ignorant l’accrétion froide,
notre étude simplifiée met en évidence l’effet du sursaut de formation d’étoiles associé à de telles fusions sur les abondances
chimiques des étoiles ainsi formées, en particulier les abondances élevées d’éléments α à métallicités élevées et les gradients verticaux d’abondances. Nous ignorons encore si l’accrétion froide seule, ou encore la dispersion de grands amas d’étoiles [55,56], peuvent reproduire de telles signatures chimiques.
Les fusions riches en gaz constituent le processus dominant
dans la formation des disques épais et fournissent une explication naturelle pour les abondances et gradients observés dans
les composantes de la Voie lactée. Une époque de fusions
riches en gaz à redshift élevé semble constituer une phase
importante dans la formation des galaxies spirales.
REMERCIEMENTS
Toutes les simulations numériques présentées dans cette article
ont été réalisées au Laboratoire d’Astrophysique Numérique de
l’Université Laval. Ces recherches sont financées par le programme des Chaires de Recherche du Canada, le Conseil de
Recherche en Science Naturelles et Génie du Canada et le
Conseil Australien de la Recherche.
BIBLIOGRAPHIE
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
D. Burstein, “Structure and origin of S0 galaxies. III - The luminosity distribution perpendicular to the plane of the disks in S0’s”, ApJ,
234, 829 (1979).
G. Gilmore et N. Reid, “New light on faint stars. III - Galactic structure towards the South Pole and the Galactic thick disc”, MNRAS,
202, 1025 (1983).
M. Chiba et T.C. Beers, “Kinematics of metal-poor stars in the Galaxy. III. Formation of the stellar halo and thick disk as revealed
from a large sample of nonkinematically selected stars”, AJ, 119, 2843 (2000).
B. Edvardsson, J. Anderson, B. Gustafsson, D.L. Lambert, P.E. Nissen et J. Tomkin, “The chemical evolution of the galactic disk –
Part one – Analysis and results”, A&A, 275, 101 (1993).
P. Girard et C. Soubiran, “[α/Fe] in the thin and the thick disc: towards an automatic parametrization of stellar spectra”, dans Threedimensional universe with Gaia, eds. C. Turon, K.S. O’Flaherty et M.A.C. Perryman (ESA SP-576), p. 169 (2005).
K. Fuhrmann, “Nearby stars of the Galactic disk and halo”, A&A, 338, 161 (1998).
T. Bensby, S. Feltzing, I. Lundström et I. Ilyin, “ α-, r-, and s-process element trends in the Galactic thin and thick disks”, A&A, 433,
185 (2005).
S. Feltzing, T. Bensby et I. Lundström, “Signatures of SN Ia in the galactic thick disk. Observational evidence from α-elements in
67 dwarf stars in the solar neighbourhood”, A&A, 397, L1 (2003).
L. Mashonkina, T. Gehren, C. Travaglio et T. Borkova, “Mg, Ba and Eu abundances in thick disk and halo stars”, A&A, 397, 275
(2003).
J.X. Prochaska, S.O. Naumov, B.W. Carney, A. McWilliam et A.M. Wolfe, “The Galactic thick disk stellar abundances”, AJ, 120, 2513
(2000).
B.E. Reddy, J. Tomkin, D.L. Lambert et C. Allende Prieto, “The chemical compositions of Galactic disc F and G dwarfs”, MNRAS,
340, 304 (2003).
K.P. Schröder et B.E.J. Pagel, “Galactic archaeology: initial mass function and depletion in the thin disc”, MNRAS, 343, 1231 (2003).
G. Tautvaisiene, B. Edvardsson, I. Tuominen et I. Ilyin, “Chemical composition of red horizontal branch stars in the thick disk of the
Galaxy”, A&A, 380, 578 (2001).
G. Gilmore, R.F.G. Wyse et J.B. Jones, “A determination of the thick disk chemical abundance distribution: Implications for galaxy
evolution”, AJ, 109, 1095 (1995).
J. Rong, R. Buser et S. Karali, “The new Basel high-latitude field star survey of the Galaxy. V. The metallicity distributions in the
inner-Galaxy fields SA 107 and NGC 6171”, A&A, 365, 431 (2001).
J.J. Dalcanton et R.A. Bernstein, ”A structural and dynamical study of late-type, edge-on galaxies. II. Vertical color gradients and the
detection of ubiquitous thick disks”, AJ, 124, 1328 (2002).
17. J. Mould, “Red thick disks of nearby galaxies”, AJ, 129, 698 (2005).
18. D. Kawata et B.K. Gibson, “GCD+: a new chemodynamical approach to modelling supernovae and chemical enrichment in elliptical
galaxies”, MNRAS, 340, 908 (2003).
19. C.B. Brook, D. Kawata, B.K. Gibson et K. Freeman, “The emergence of the thick disk in a cold dark matter universe”, ApJ, 612, 894
(2004)
20. T.J. Davidge, “The disk and extraplanar regions of NGC 55”, ApJ, 622, 279 (2005).
21. T. Padmanabhan, Structure Formation in the Universe (Cambridge: Cambridge University Press) (1993).
22. C. Power, Thèse de doctorat, University of Durham (2003).
23. A.C. Quillen et D. Garnett, “The saturation of disk heating in the solar neighborhood and evidence for a merger 9 Gyr ago”, dans
Galaxy Disks and Disk Galaxies, ASP Conference Series 230, eds. G. Jose, S.J. Funes et E.M. Corsini, p. 87 (2001).
24. B. Nordström, et al., “The Geneva-Copenhagen survey of the Solar neighbourhood. Ages, metallicities, and kinematic properties of
14000 F and G dwarfs”, A&A, 418, 989 (2004).
25. C.B. Brook, D. Kawata, B.K. Gibson et C. Flynn, “Stellar halo constraints on simulated late-type galaxies”, MNRAS, 349, 52 (2004).
26. A. Meza, J.F. Navarro, G. Abadi et M. Steinmetz, “Accretion relics in the solar neighbourhood: debris from ω Cen’s parent galaxy”,
MNRAS, 359, 93 (2005).
27. S.E. Woosley et T.A. Weaver, “The evolution and explosion of massive stars. II. explosive hydrodynamics and nucleosynthesis”, ApJS,
101, 181 (1995).
28. G. Gilmore, R.F.G. Wyse et K. Kuijken, “Kinematics, chemistry, and structure of the Galaxy”, ARA&A, 27, 555 (1989).
29. R.F.G. Wyse, “Galactic Structure”, dans The Local Group as an Astrophysical Laboratory, ed. M. Livio (Cambridge: Cambridge
University Press) (2004).
30. C. Murali, N. Katz, L. Hernquist, D.H. Weinberg et R. Davé, “The growth of galaxies in cosmological simulations of structure formation”, ApJ, 571, 1 (2002).
31. A.C. Seth, J.J. Dalcanton et R.S. de Jong, “A study of edge-on galaxies with the Hubble Space Telescope Advanced Camera for
Surveys. I. Initial results”, AJ, 129, 1331 (2005).
32. B. Robertson, J.S. Bullock, A.S. Font, K.S. Johnston et L. Hernquist, “ Cold Dark Matter, stellar feedback, and the galactic halo abundance pattern”, ApJ, 632, 872 (2005).
33. P.J. Quinn, L. Hernquist et D.P. Fullagar, “Heating of galactic disks by mergers”, ApJ, 403, 74 (1993).
34. C.B. Brook, D. Kawata, B.K. Gibson et C. Flynn, “Galactic halo stars in phase space: A hint of satellite accretion?”, ApJ, 585, L125
(2003).
35. N.F. Martin, R.A. Ibata, M. Ballazzini, M.J. Irwin, G.F. Lewis et W. Dehnen, “A dwarf galaxy remnant in Canis Major: the fossil of an
in-plane accretion on to the Milky Way”, MNRAS, 348, 12 (2004).
36. J.F. Navarro, A. Helmi et K.C. Freeman, “The extragalactic origin of the Arcturus group”, ApJ, 601, L43 (2004).
37. B. Yanny et al., “A low-latitude halo stream around the Milky Way”, ApJ, 588, 824 (2003).
38. A. Helmi, S.D.M. White, P.T. de Zeeuw et H. Zhoa, “Debris streams in the solar neighbourhood as relicts from the formation of the
Milky Way”, Nature, 402, 53 (1999).
39. R.A. Ibata, G. Gilmore et M.J. Irwin, “A dwarf satellite galaxy in Sagittarius,” Nature, 370, 194 (1994).
40. P. Yoachim et J.J. Dalcanton, “Structural parameters of thin and thick disks in edge-on disk galaxies”, AJ, 131, 226 (2006).
41. V. Springel et L. Hernquist, “Formation of a spiral galaxy in a major merger”, ApJ, 622, L9 (2005).
42. B.E. Robertson, J.S. Bullock, T.J. Cox, T. Di Matteo, L. Hernquist, V. Springel et N. Yoshida, “A merger-driven scenario for cosmological disk galaxy formation”, ApJ, 645, 986 (2006).
43. J.J. Dalcanton et R.A. Bernstein, “A structural and dynamical study of late-type, edge-on galaxies. I. Sample selection and imaging
data”, AJ, 120, 203 (2000).
44. C.B. Brook, B.K. Gibson, H. Martel et D. Kawata, “The emergence of the thick disk in a CDM universe. II. Colors and abundance
patterns”, ApJ, 630, 298 (2005).
45. A.S. Font, K.V. Johnston, J.S. Bullock et B.E. Robertson, “Chemical abundance distributions of galactic halos and their satellite systems in a CDM universe”, ApJ, 638, 585 (2006).
46. A. Renda, B.K. Gibson, M. Mouhcine, R.A. Ibata, D. Kawata, C. Flynn et C.B. Brook, “The stellar halo metallicity-luminosity relationship for spiral galaxies”, MNRAS, 363, L16 (2005).
47. F. Governato, B. Willman, L. Mayer, A. Brooks, G. Stinson, O. Valenzuela, J. Wadsley et T. Quinn, “Forming disc galaxies in Λ CDM
simulations”, MNRAS, 374, 1479 (2006).
48. B.E. Reddy, D.L. Lambert et P.C. Allende, “Elemental abundance survey of the Galactic thick disc”, MNRAS, 367, 1329 (2006).
49. K. Kuijken et J. Dubinski, “Nearly self-consistent disc/bulge/halo models for galaxies”, MNRAS, 277, 1341 (1995).
50. R.J. Lavery, A. Remijan, V. Charmandaris, R.D. Hayes et A.A. Ring, “Probing the evolution of the galaxy interaction/merger rate
using collisional ring galaxies”, ApJ, 612, 679 (2004).
51. C.B. Brook, D. Kawata, H. Martel, B.K. Gibson et J. Bailin, “Disk evolution since z ~ 1 in a CDM Universe”, ApJ, 639, 126 (2006).
52. J.M. Schombert, S.S. McGaugh et J.-A. Eder, “Gas mass fractions and the evolution of low surface brightness dwarf galaxies”, AJ,
121, 2420 (2001).
53. D.K. Erb, A.E. Shapley, M. Pettini, C.C. Steidel, N.A. Reddy et K. L. Adelberger, “The mass-metallicity relation at z ~ 2”, ApJ, 644,
813 (2006).
54. J.J. Dalcanton, “The metallicity of Galaxy disks: infall versus outflow”, ApJ, 658, 941 (2007).
55. B.G. Elmegreen et D.M. Elmegreen, “Observations of thick disks in the Hubble Space Telescope Ultra Deep Field”, ApJ, 650, 644
(2006).
56. P. Kroupa, “Thickening of galactic discs through clustered star formation”, MNRAS, 330, 707 (2002).
244 C PHYSICS
IN
ARTICLE DE FOND
RESULTS
BY
FROM THE
GEMINI DEEP DEEP SURVEY
R.G. ABRAHAM, I. DAMJANOV, E. MENTUCH, P. NAIR, R. CARLBERG, D. CRAMPTON,
R. MUROWINSKI, K. GLAZEBROOK, P. MCCARTHY, H. YAN, S. SAVAGLIO, D. LE BORGNE,
H-W. CHEN, I. JØRGENSEN, K. ROTH, S. JUNEAU, AND R. MARZKE
T
his article presents some highlights from the first
ten papers produced by the Gemini Deep Deep
Survey (GDDS) [1]. The goal of the GDDS is to
use a stellar-mass-selected sample to probe
galaxy evolution at redshifts, z, in the interval 1 < z < 2.
This range of redshifts corresponds to times when the
Universe is between three and six billion years old. This is
a special time in the history of the Cosmos, because it
spans the epoch at which the rate at which stars form in the
Universe is at its peak, and over which around half of all
the existing stars in the Universe formed. Sadly, it also
corresponds to a time where relatively little is known
about galaxies. In fact, at the time the GDDS was proposed, galaxies with redshifts in the range 1 < z < 2 were
so rare that this interval in cosmic history was known as
the ‘redshift desert’.
from which one can obtain a redshift using spectrographs
which operate at visible wavelengths. It is possible to
obtain redshifts using near-infrared spectrographs, but at
present these are inefficient compared to visible wavelength spectrographs, and near-infrared spectroscopy has
only yielded redshifts for a handful of objects. It is possible to forego the need for emission lines completely and to
obtain redshifts using the overall shape of a galaxy’s spectrum (the so-called continuum spectrum), but this requires
very high signal-to-noise spectroscopy, which is difficult
to obtain for faint galaxies, because the night sky changes
in brightness (at a low level) on a timescale of minutes.
This changing sky brightness results in imperfect sky
removal and limits the effective exposure time to a few
hours with most spectrographs. Fortunately, these prob-
Why is it so difficult to obtain redshifts for galaxies in this
particular redshift range? The main reason is that most
galaxy redshifts are obtained using bright emission lines.
The majority of these lines are produced at wavelengths
which cluster at either visible wave-lengths, or else at farultraviolet wavelengths, with few lines in between (in the
near-ultraviolet). As galaxies are redshifted, visible wavelength emission is shifted into the near-infrared, and the
gap in the near-ultraviolet is redshifted into visible wavelengths. Thus, by the time we reach z=1, few lines remain
SUMMARY
The Gemini Deep Deep Survey is a survey of
galaxies in the redshift range 1 < z < 2 whose
main purpose is to determine the abundance
of galaxies as a function of mass at the time
when the Universe was forming stars most
quickly. In a series of papers published over
the last four years, the survey has shown
that massive and old galaxies are surprisingly common in the distant Universe, lending
strong support to a new paradigm for galaxy
formation known as 'Cosmic Downsizing'. In
many ways the oldest and most massive
objects in the survey resemble nearby elliptical galaxies, but they also show some rather
interesting differences, such as being much
more compact and dense than nearby galaxies of similar mass.
Fig. 1
Spectra of evolved/quiescent GDDS galaxies with z >
1.3. The SDSS Lumi-nous Red Galaxy composite has
been overlaid on each spectrum and an offset has been
applied to each spectrum in order to stack them vertically. The locations of the stellar MgII2800 and MgI2852
lines are indicated by the dashed lines. Taken from
GDDS paper IV [2].
R.G. Abraham
<abraham@astro.
utoronto.ca>,
I. Damjanov,
E. Mentuch, P. Nair,
R. Carlberg (University
of Toronto);
D. Crampton,
R. Murowinski (Herzberg
Institute of Astrophysics,
National Research
Council of Canada);
K. Glazebrook
(Swinburne University of
Technology);
P. McCarthy, H. Yan
(Observatories of the
Carnegie Institution of
Washington);
S. Savaglio (MaxPlanck-Institut fur
extraterrestrische
Physik);
D. Le Borgne (Institut
d’Astro-physique de
Paris);
H-W. Chen (University
of Chicago);
I. Jørgensen, K. Roth
(Gemini Observatory);
S. Juneau (University of
Arizona);
R. Marzke (San
Francisco State
University)
RESULTS FROM THE GDDS (ABRAHAM)
lems go away when studying the most distant galaxies, because
these are at z > 3 and for these objects prominent far-UV spectral lines are once again redshifted into visibility, so paradoxically it is easier to obtain redshifts for these very distant galaxies than it is to obtain redshifts in the ‘redshift desert’.
By the time the GDDS had begun some progress had been
made in overcoming these limitations, and decent-sized samples of galaxies at 1 < z < 2 were at last being obtained.
However, these samples were restricted to young galaxies,
which are very luminous in the near-ultraviolet, and older
galaxies (assuming they exist) would be missed by this technique. Older galaxies in the distant Universe are arguably the
most interesting to understand, since they form soon after the
big bang, and could well contain most of the stars. Our ambition with the GDDS was to look for these older galaxies in the
redshift desert. To do this, we needed a new technique that
would allow us to expose on small patches of sky for times
much longer than is possible using conventional spectroscopy,
so that we could obtain redshifts for these galaxies using their
continuum spectra. As has been noted, conventional spectrographs do not allow these sorts of observations to be done
effectively because of sky subtraction limitations, so we needed a better approach.
Fig. 2
The universal star-formation rate per unit volume for galaxies in different stellar mass ranges presented in GDDS Paper
VI [3]. Note how the star-formation rate evolution is a very
strong function of stellar mass. These star-formation rates
were derived from L([OII]) (circles) and from rest-UV flux
(triangles). The symbols are color-coded by the logarithmic
mass ranges labeled in the figure. The error bars in redshift
show the width of the redshift bins used. The squares are the
values found locally by Brinchmann et al. [4] converted
according to our assumed IMF and dust correction. The compilation made by Hopkins [5], where all the values are converted to a (ΩM = 0.3, ΩΛ = 0.7, h = 0.7) cosmology, are overplotted with diamonds. The line is the fit derived by Cole
et al. [6] assuming AV = 0.6 (solid line).
246 C PHYSICS
IN
Our strategy was to implement an innovative new approach to
sky subtraction and multiplexing on the Gemini telescopes.
This technique, known as “nod and shuffle”, was proposed
independently by Glazebrook & Bland-Hawthorn in 2001 [7],
and (in a less-developed form) by Cuillandre et al. in 1994 [8].
The idea is to use part of the CCD detector as a “storage register” for a beam-switched image. Beam-switching is achieved
by rapid alternation between object and sky positions (“nodding”), which is undertaken with no detector readout penalty
(because modern detectors allow one to move charge around
without reading out the device). Instead, the sky image is shuffled to a storage region on the CCD. Typically, nodding takes
place every 30 to 60 seconds, which is a timescale faster than
the variations of the airglow emission lines that dominate the
sky background. Because both the sky and objects are observed
quasi-simultaneously through the same optical path, slits and
pixels, nod and shuffle results in an order of magnitude
improvement in sky subtraction, opening up significant new
observational capabilities for large telescopes. For example,
very deep integrations (10 times longer than is practical with
conventional spectroscopy) are possible with nod and shuffle.
Using nod & shuffle on the Gemini telescope, integration times
as long as 30 hours per field allowed us to determine continuum redshifts from rest-frame UV metallic absorption features,
rather than relying on emission lines from star-forming galaxies (see figure 1).
The main result from the GDDS is that massive old galaxies
are far more common than originally expected at z ~ 1.5
(GDDS Paper III), and that many of these systems are surprisingly old (GDDS Paper IV). It is important to note though that
similar goal to ours motivated the K20 Survey [9] on the VLT
(an ESO Key Project), and the two surveys reported similar
results at similar times. Amusingly, key papers even appeared
in the same issue of Nature [10,11]. For example, GDDS and
K20 obtained similar evolving stellar mass functions and ages
for massive ‘red and dead’ systems. In retrospect, the convergence on such fundamental results from two independent surveys strengthened the credibility of the results from both teams.
Taken together, these early results lent considerable impetus to
Cowie et al.’s notion of galactic ‘downsizing’, evidence for
which was summarized in GDDS Paper V [12]. In this ‘downsizing’ picture, the most massive galaxies form first (soon after
the big bang), and as the Universe evolves less massive galaxies form (see figure 2). This picture was in direct contradiction
with popular hierarchical models where galaxy assembly
traced the build up of dark matter haloes, although recent versions of these models now are able to account for downsizing
by decoupling the behaviour of galaxies from that of the underlying dark matter. At least partly a result of observations by the
GDDS, K20 (and other surveys), it is probably fair to say that
‘downsizing’ is now entrenched as the defacto observational
paradigm for high-redshift galaxy formation.
More recent papers in the GDDS series have tended to use the
survey data as a starting point for analyses of data obtained
from other facilities (e.g. the Hubble Space Telescope and the
Spitzer Space Telescope). Data from the Hubble Space
RESULTS FROM THE GDDS (ABRAHAM)AA
Fig. 4
Near-infrared images obtained with the Hubble Space
Telescope’s NICMOS camera of GDDS galaxies, as presented in GDDS Paper X [14]. The three columns show the galaxies (left column), the best fitting model (middle column), and
residuals obtained by subtracting the models from the galaxies (right column). The residual images have been scaled by
a factor of 10 compared to the data and models to bring out
faint features. The bars at the bottom are one arcsecond in
length. Note the very compact nature of several of the systems shown.
Postage-stamp images taken from GDDS Paper VIII [13]
showing the mor-phologies of the 54 galaxies in our GDDS
sample with log(stellar mass) > 10.5. These galaxies have
been sorted in order of decreasing redshift. Early-type galaxies are circled. Each image is 5 arcsec by 5 arcsec in size, and
labeled with the galaxy's ID number, spectroscopic classification, redshift confidence class, rest-frame (U-B) color, redshift, and stellar mass. Objects without high-confidence spectroscopic redshifts have their redshifts labeled in parentheses.
The border of each galaxy image is colored according to the
galaxy's spectroscopic classification. Objects with red borders have evolved spectra. The gray regions surrounding
groups of postage stamps indicate which of three broad redshift bins the objects fall within.
change their appearance by simply rearranging populations of
existing stars, and the peak era for structural change in galaxies does not necessarily correspond to the peak era for star formation. This is also seen at higher redshifts than those probed
by the GDDS, where the paucity of evolving red galaxies at
z > 2 in deep infrared samples [15-17], shows that the assembly
epoch for elliptical galaxies is probably decoupled from the
epoch at which most of the stars in the galaxy formed.
Telescope has proved particularly valuable, forming a key
component of the analysis in five of the later GDDS papers.
For example, in GDDS Paper VIII we used the Hubble Space
Telescope's Advanced Camera for Surveys to measure the mass
density function of morphologically selected early-type galaxies in the GDDS fields. We find that at z = 1 approximately
70% of the stars in massive galaxies live in elliptical galaxies
(figure 3). This fraction is remarkably similar to that seen in the
local Universe. However, we also detect very rapid evolution in
the abundance of massive red elliptical galaxies over the range
1.0 < z < 1.6, suggesting that in this epoch the strong colormorphology relationship seen in the nearby Universe is beginning to fall into place. This works begins to place downsizing
in a broader context which encompasses the formation of the
standard Hubble sequence. More importantly, it also shows that
the space density of an established galaxy class (elliptical
galaxies) can evolve strongly, even as the stars within the
galaxy evolve weakly. One must therefore be careful to decouple structural evolution of galaxies from the evolution of the
stars within the galaxies. It seems that even old galaxies can
An extreme example of this phenomenon may well have
emerged from the most recent GDDS paper [14], which analyzed the results of near-infrared imaging of 16 high mass
passively evolving galaxies, most of which were old (ages >
1 Gyr). Most of these galaxies show compact regular morphologies consistent with classical elliptical galaxies.
However, around one-third of these galaxies are extraordinarily compact, even though they are massive (figure 4). These
elliptical galaxies are a factor of 2–3 smaller than elliptical
galaxies of similar mass today. While similar systems have
been seen at z>2 [18], detection of old counterparts to these
objects in the GDDS shows that these galaxies must somehow
change their size quite radically after their stars are already
mature. Similarly compact massive galaxies are completely
absent in the nearby Universe, and the objects seen in the
GDDS study have mass densities that are an order of magnitude larger then elliptical galaxies today. Damjanov et al. [14]
also show that size evolution occurs primarily in the 1.1 < z <
1.5 redshift interval, or over a time of only 1.6 billion years.
The galaxies seen by Damjanov et al. [14] are already as mas-
Fig. 3
RESULTS FROM THE GDDS (ABRAHAM)
most interesting science to emerge from
the GDDS has nothing to do with the
original science goals of the survey. A
prime example of this is the work that has
emerged on the metallicities of
0.4 < z < 1 galaxies. GDDS Paper VII by
Savagio et al. [22] reports the discovery of
a mass-metallicity (MZ) relation in high
redshift galaxies. This result emerges
from an investigation of the bluer galaxies on our spectroscopic masks. The standard way to determine metallicities in
high-z galaxies is to use emission lines
originating in HII regions. By definition,
this technique is selecting young star
forming galaxies which, as noted earlier,
are not the targets of priority for the
GDDS. However, we did have significant
numbers of these systems in our survey as
second priority targets. By combining
spectra for 29 blue GDDS galaxies with a
similar data set from the CFHT at lower
redshift (obtained as part of the Canada
France Redshift Survey; CFRS [23] we
Fig. 5 Effective radius Re as a function of stellar mass for five samples of early-type galaxies obtained a total sample of 65 galaxies.
in the redshift range 1.1 < z < 2. Points are color-coded by two redshift ranges (red = The resulting mass-metallicity relation,
z>1.46, blue = z < 1.46). Different symbols correspond to different surveys. The size- clearly detected in the GDDS and CFRS
mass relation for local early-type galaxies in the Sloan Digital Sky Survey is presented sample, is different from the same relawith sizes taken from Bernardi et al. [19], and matched with masses calculated following tion defined by Tremonti et al. in the local
Baldry et al. [20] (black points). Contours represent linearly spaced regions of constant universe [24]. It appears that by z ~ 0.7
density of galaxies in size-mass parameter space. The solid line is the best-fit relation to massive galaxies have achieved a mature
the data points at redshifts 1.2 < z < 2. Note that at high redshift elliptical galaxies of a
chemical state, similar to massive galaxgiven mass are a factor of 2-3 smaller than their counterparts at low redshift. Three
arrows denote the effects of various models for size growth. The correspond to equal ies at z ~ 0.1. On the other hand, small
mass dry mergers [21], adiabatic ex-pansion with 50% mass loss, and pure size evolution galaxies are still in the process of forming
at constant stellar mass. The arrows denote the effects of these on the positions of both their metals. By combining our massmetallicity relation with the local relathe least and the most massive galaxy.
tion, we were able to build up a simple
empirical model for the evolution in the
relation which reproduces the more recent mass-metallicity
sive as the most massive field galaxies seen nearby (figure 5),
relation in z ~ 3 Lyman break galaxies reported by Erb et
suggesting that they must somehow grow bigger without growal. [25]. This model is also consistent with the downsizing
ing more massive. At present no physical mechanism for growdescribed earlier, since in our model to the e-folding time for
ing a dense, massive, old galaxy by a large factor without formstar formation is inversely proportional to the initial mass of
ing large numbers of new stars is known, so it seems that there
galaxies, and it provides an extra dimension to the exploration
must be a missing ingredient in our understanding of galaxy
of galaxy formation that was completely unanticipated when
formation.
we proposed the survey: investigation of the time evolution of
the heavy element enrichment in galaxies with different initial
Although the highest-impact papers to have emerged (so far)
conditions.
from the GDDS focus on the abundance of massive red galaxies at high redshift, it comes as no surprise that much of the
REFERENCES
1.
2.
3.
Abraham, R.G., et al., “The Gemini Deep Deep Survey. I. Introduction to the Survey, Catalogs, and Composite Spectra” (Paper I), AJ,
127, 2455, (2004).
McCarthy, P.J., et al., “Evolved Galaxies at z > 1.5 from the Gemini Deep Deep Survey: The Formation Epoch of Massive Stellar
Systems” (Paper IV), ApJ, 614L, 9 (2004).
Juneau, S. et al., “Cosmic Star Formation History and Its Dependence on Galaxy Stellar Mass” (Paper V), ApJ, 619L,135 (2005).
248 C PHYSICS
IN
RESULTS FROM THE GDDS (ABRAHAM)AA
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
19.
20.
21.
22.
23.
24.
25.
Brinchmann, J., Charlot, S., White, S.D.M., Tremonti, C., Kauffmann, G., Heckman, T., & Brinkmann, J., MNRAS, 351, 1151 (2004).
Hopkins, A.M., ApJ, 615, 209 (2004).
Cole, S., et al., MNRAS, 326, 255 (2001).
Glazebrook, K. and Bland-Hawthorn, J., PASP, 113, 197 (2001).
Cuillandre, J.C., Fort, B., Picat, J.P., Soucail, J.P., Altieri, B., Beigbeder, F., Duplin, J.P., Pourthie, T., & Ratier, G., A&A, 281, 603
(1994).
Cimatti, A., et al., A&A, 482, 21 (2008).
Glazebrook, K., et al., “A high abundance of massive galaxies 3-6 billion years after the Big Bang” (Paper III), Nature, 430, 181
(2004).
Fontana, A., et al., A&A, 424, 23 (2004).
Cowie, L.L., Songaila, A., Hu, E.M. & Cohen, J.G., AJ, 112, 839 (1996).
Abraham, R.G. et al., “The Gemini Deep Deep Survey: VIII. When Did Early-type Galaxies Form?” (Paper VIII), ApJ, 669, 184
(2007).
Damjanov et al., “Red Nuggets at z ~ 1.5: Compact passive galaxies and the formation of the Kormendy Relation” (Paper X), ApJ,
submitted, (2008).
Kriek, M., et al., ApJ, 649, L71 (2006).
Labbé, I., et al., ApJ, 624, L81 (2005).
Cimatti, A., et al., Nature, 430, 184 (2004).
van Dokkum, P.G., et al., ApJ, 677, L5 (2008).
Bernardi, M., et al., AJ, 125, 1817 (2003).
Baldry, I.K., Glazebrook, K., & Driver, S.P. , MNRAS, 388, 945 (2008).
Boylan-Kolchin, M., Ma, C., & Quataert, E., MNRAS, 369, 1081 (2006).
Savaglio, S., et al., “The Gemini Deep Deep Survey. VII. The Redshift Evolution of the Mass-Metallicity Relation” (Paper VII), ApJ,
635, 260 (2005).
Lilly, S.J., Le Fevre, O., Hammer, F., & Crampton, D. ApJ, 460L, 1 (1996).
Tremonti, C.A., Heckman, T.M., Kauffmann, G., Brinchmann, J., Charlot, S., White, S.D.M., Seibert, M., Peng, E.W., Schlegel, D.J.,
Uomoto, A., Fukugita, M., & Brinkmann, J., ApJ, 613, 898 (2004).
Erb, D.K., Shapley, A.E., Pettini, M., Steidel, C.C., Reddy, N.A., & Adelberger, K.L., ApJ, 644, 813 (2006).
ADDITIONAL REFERENCES
Kriek, M., et al., ApJ, 669, 776 (2007).
Le Borgne, D., et al., “Gemini Deep Deep Survey. VI. Massive Hästrong Galaxies at z ~ 1” (Paper VI), ApJ, 642, 48 (2006).
Longhetti, M., et al., MNRAS, 374, 614 (2007).
McCarthy, P.J., et al., “A Compact Cluster of Massive Red Galaxies at a Redshift of 1.5” (Paper IX), ApJ, 664, 17 (2007).
McGrath, E.J., Stockton, A. & Canalizo, G., ApJ, 669, 241 (2007).
Savaglio, S., et al., “The Gemini Deep Deep Survey. II. Metals in Star-forming Galaxies at Redshift 1.3 < z < 2” (Paper II), ApJ, 602, 51
(2004).
DEPARTMENTAL MEMBERS / MEMBRES DÉPARTEMENTAUX
- Physics Departments / Départements de physique (as at 2008 November 1 / au 1er novembre 2008)
Acadia University
Bishop's University
Brandon University
Brock University
Carleton University
Collège Ahuntsic
Collège François-Xavier-Garneau
Collège Montmorency
Concordia University
Dalhousie University
École Polytechnique de Montréal
Lakehead University
Laurentian University
McGill University
McMaster University
Memorial Univ. of Newfoundland
Mount Allison University
Okanagan University College
Queen's University
Royal Military College of Canada
Saint Mary’s University
Simon Fraser University
St. Francis Xavier University
Trent University
Université de Moncton
Université de Montréal
Université de Sherbrooke
Université du Québec à Trois-Rivières
Université Laval
University of Alberta
University of British Columbia
University of Calgary
University of Guelph
University of Lethbridge
University of Manitoba
University of New Brunswick
University of Northern British Columbia
University of Ontario Inst. of Technology
University of Ottawa
University of Prince Edward Island
University of Regina
University of Saskatchewan (and Eng. Phys.)
University of Toronto
University of Toronto (Medical Biophysics)
University of Victoria
University of Waterloo
University of Western Ontario
University of Windsor
University of Winnipeg
Wilfrid Laurier University
York University
SUSTAINING MEMBERS / MEMBRES DE SOUTIEN
(as at 2008 November 1 / au 1er novembre 2008)
Richard Hemingway
Akira Hirose
Thomas Jackman
Thayyil Jayachandran
Béla Joós
James D. King
Ron M. Lees
Louis Marchildon
David B. McLay
Jean-Louis Meunier
J.C. Douglas Milton
Michael Morrow
A. John Alcock
Thomas K. Alexander
C. Bruce Bigham
Harvey Buckmaster
Allan I. Carswell
Walter Davidson
M. Christian Demers
Fergus Devereaux
Marie D'Iorio
Gerald Dolling
Gordon W.F. Drake
Elmer H. Hara
Michael Kevin O'Neill
Allan Offenberger
A. Okazaki
Shelley Page
Roger Phillips
Beverly Robertson
Robert G.H. Robertson
Pekka Sinervo
Boris P. Stoicheff
Eric C. Svensson
Louis Taillefer
John G.V. Taylor
Andrej Tenne-Sens
Michael Thewalt
Greg J. Trayling
William Trischuk
Sreeram Valluri
Henry M. Van Driel
Paul S. Vincett
Erich Vogt
Andreas T. Warburton
CORPORATE-INSTITUTIONAL MEMBERS /
MEMBRES CORPORATIFS-INSTITUTIONNELS
(as at 2008 November 1 / au 1er novembre 2008)
The Corporate and Institutional Members of the Canadian Association
of Physicists are groups of corporations, laboratories, and institutions
who, through their membership, support the activities of the Association. The entire proceeds of corporate membership contributions are
paid into the CAP Educational Trust Fund and are tax deductible.
CORPORATE / CORPORATIFS
BUBBLE TECHNOLOGY INDUSTRIES
CANADA ANALYTICAL & PROCESS TECH.
CANBERRA CO.
E-INSTRUCTION
Les membres corporatifs et institutionnels de l'Association canadienne des
physiciens et physiciennes sont des groupes de corporations, de laboratoires
ou d'institutions qui supportent financièrement les activités de l'Association.
Les revenus des contributions déductibles aux fins d'impôt des membres
corporatifs sont entièrement versés au Fonds Educatif de l'ACP.
GLASSMAN HIGH VOLTAGE INC.
JOHNSEN ULTRAVAC INC.
KURT J. LESKER CANADA INC.
PLASMIONIQUE INC.
VARIAN CANADA INC.
The Canadian Association of Physicists cordially invites interested corporations and institutions to make application for Corporate or Institutional
membership. Address all inquiries to the Executive Director.
INSTITUTIONAL / INSTITUTIONNELS
ATOMIC ENERGY OF CANADA LIMITED
CANADIAN LIGHT SOURCE
INSTITUTE FOR QUANTUM COMPUTING
PERIMETER INSTITUTE FOR THEORETICAL PHYSICS
TRIUMF
L'Association canadienne des physiciens et physiciennes invite cordialement corporations et institutions à faire partie des membres corporatifs
ou institutionnels. Renseignements auprès de la directrice exécutive.
CANADIAN ASSOCATION OF PHYSICISTS / ASSOCIATION CANADIENNE DES PHYSICIENS ET PHYSICIENNES
Bur. Pièce 112, Imm. McDonald Bldg., Univ. of/d’Ottawa, 150 Louis Pasteur, Ottawa, Ontario K1N 6N5
Phone / Tél : (613) 562-5614; Fax / Téléc : (613) 562-5615 ; Email / courriel : [email protected]
INTERNET - HTTP://WWW.CAP.CA
250 C PHYSICS
IN
ARTICLE DE FOND
LES
GALAXIES À SURSAUTS DE FORMATION STELLAIRE
DANS L’ULTRAVIOLET
PAR
CARMELLE ROBERT
L
es galaxies à sursauts de formation stellaire,
communément appelées « starbursts », produisent
des étoiles avec une intensité de formation stellaire qui peut atteindre 1 à 100 M an-1 kpc-2 [1].
Cette activité exceptionnelle excède par un facteur 10 à
1000 le taux de formation stellaire d’une galaxie spirale
normale comme la Voie lactée et ne saurait durer qu’un
temps limité comme le dicte le réservoir de gaz
disponible.
Les étoiles massives (M$10 M ), qui ont une vie relativement courte, sont observées en grand nombre dans les
starbursts. Elles sont en fait responsables des grandes
luminosités des sursauts dans la plupart des domaines de
longueur d’onde. Les étoiles massives, étant des objets
chauds, ont leur maximum d’émission dans l’ultraviolet
(UV). Une grande portion de cette radiation UV n’est pas
observée (selon Buat et al. [2], seulement 33 % de l’émission UV s’échappe pour être observée), mais ionise le
milieu interstellaire qui, par recombinaison, produit des
raies spectrales et un continuum dans le visible et l’infrarouge (IR) proche. La poussière, grandement présente
dans les sites de formation stellaire jeune, absorbe aussi
une partie de l’émission ultraviolette des étoiles massives
pour produire une émission thermique observable dans
l’IR lointain. Les étoiles massives ont aussi un impact
important sur leur environnement. Par le biais de leurs
vents et de leur explosion en supernova, elles injectent de
grandes quantités d’énergie mécanique dans le milieu
RÉSUMÉ
Les « starbursts », sursauts de formation
stellaire intense et violente, sont très
fréquents partout dans l’Univers et représentent une phase importante de l’évolution des
galaxies. Dans le domaine des longueurs
d’onde de l’ultraviolet, on retrouve des signatures directes des étoiles massives qui
offrent un avantage unique pour la caractérisation du contenu stellaire, de la fonction de
masse initiale et du mode de formation stellaire des sursauts. Cet article présente la
technique de synthèse spectrale de l’ultraviolet utilisée pour décrire les sursauts et
résume les grandes conclusions obtenues
de son application.
interstellaire. Elles représentent aussi la source principale
d’enrichissement en éléments lourds, définissant la métallicité des étoiles à venir.
Les starbursts constituent une composante importante de
notre Univers. On compte quatre grandes galaxies à sursauts (M82, NGC253, NGC4945 et M83) dans un rayon
de 10 Mpc (30 millions d’années-lumière) autour de nous.
À elles seules, ces galaxies sont responsables de 25 % du
taux de formation stellaire local [3]. L’ampleur des sursauts
est très variable, allant des régions HII géantes, comme 30
Dor dans le Grand Nuage de Magellan [4] au « Lyman
Break Galaxies » à un grand décalage spectral (z $ 2) [5],
en passant par les petites galaxies irrégulières bleues,
comme IZw18 [6], les starbursts nucléaires, comme le prototype NGC7714 [7] et les galaxies ultra-lumineuses IR
découvertes par IRAS [8]. Certains chercheurs préfèrent
distinguer galaxies starbursts et régions starbursts en se
basant sur l’importance de la luminosité des sursauts par
rapport à la galaxie hôte. Néanmoins, une région starburst
se limite généralement à un amas compact, dont le
diamètre atteint 10 à 100 pc, et renferme entre 104 et
107 M sous forme d’étoiles [9]. Plusieurs de ces régions
sont présentes pour former une galaxie starburst comme
on a pu l’observer sur les premières images ultraviolettes
détaillées obtenues avec le télescope spatial Hubble. Ces
mêmes images montrent aussi une émission diffuse importante (75 % du flux UV total du starburst) sous-jacente aux
amas de formation stellaire. La dissipation spatiale d’amas
plus âgés [10] est l’une des hypothèses proposées pour
expliquer cette émission diffuse.
L’origine des starbursts est intimement reliée à des perturbations de la composante gazeuse des galaxies. Une corrélation existe entre la présence des sursauts et les signatures de collisions et des effets de marée observés dans
plusieurs galaxies [11]. Dans les galaxies starbursts plus
isolées, des processus internes sont évoqués pour expliquer la compression du gaz. Parmi ces processus, on propose des instabilités associées à une barre [12] ou à l’activité d’un noyau central [13]. Des sursauts séquentiels,
causés par l’énergie mécanique des vents stellaires et des
supernovae d’un premier événement, pourraient aussi
jouer un rôle sur l’ampleur du phénomène.
Carmelle Robert
<[email protected].
ca>, Département de
physique, de génie
physique et d’optique,
et Centre de
recherche en astrophysique du Québec,
Université Laval,
Québec, QC G1K
7P4
Importance des observations ultraviolettes
L’observation ultraviolette est très sensible à la présence
d’étoiles massives, i.e. de type OB, et à leurs descendantes
LES GALAXIES À SURSAUTS ... (ROBERT)
évolutives, les étoiles de type Wolf-Rayet. La distribution de
l’énergie dans l’ultraviolet change énormément entre une population stellaire jeune et une population d’âge intermédiaire
(~1 Ga) [14]. Ce domaine de longueurs d’onde permet d’établir
l’âge d’une population, mais aussi d’étudier la fonction de
masse initiale (FMI), la composition chimique, ainsi que le
mode et le taux de formation des étoiles. Ces paramètres
représentent des indices importants pour la compréhension de
la formation stellaire et de l’évolution des galaxies.
L’atmosphère terrestre bloquant les photons ultraviolets,
l’étude des starbursts a connu un essor important avec les missions spatiales. Kinney et al. [15] ont produit le premier atlas de
spectres de galaxies starbursts entre 1200 et 3300 Å avec des
données de l’IUE (International Ultraviolet Explorer). Le télescope spatial Hubble a été le premier à offrir des images et
spectres ultraviolets d’une résolution spatiale exceptionnelle
(e.g. Meurer et al. [9], Whitmore et al. [16] et Leitherer et
al. [17]). Les missions HUT (Hopkins Ultraviolet Telescope) et
FUSE (Far-Ultraviolet Spectroscopic Explorer, supporté par
l’Agence spatiale canadienne) ont grandement ajouté à la collection de spectres en atteignant l’ultraviolet lointain sous les
1200 Å (e.g. Gonzalez-Delgado et al. [18] et Pellerin &
Robert [19]). Les images du satellite GALEX, avec un champ
de vision encore plus important, permettent une étude systématique incomparable des galaxies starbursts proches
(e.g. Bianchi et al. [20]).
Cet article discute de l’importance de la spectroscopie ultraviolette pour l’étude du contenu stellaire dans les starbursts.
Ces travaux sont grandement motivés par le fait que les starbursts représentent des objets clefs pour la compréhension de
notre Univers. Ils sculptent l’allure des galaxies et marquent
leur évolution. Ils constituent de plus des laboratoires idéaux
pour l’étude des étoiles massives et de leur environnement.
SPECTROSCOPIE ULTRAVIOLETTE ET
CONTENU STELLAIRE
Dans l’ultraviolet, le spectre d’un starburst montre des raies
d’absorption qui prennent forme dans le vent et la photosphère
des étoiles OB ainsi que dans le milieu interstellaire.
Contrairement aux raies stellaires présentes dans le domaine du
visible, celles de l’ultraviolet ne sont pas cachées par des raies
d’émission du gaz qui accompagne la formation stellaire.
Les raies UV qui marquent particulièrement bien la présence
d’étoiles OB dans une population stellaire jeune sont celles de
PV λλ1118,1128, SiIV λλ1122,1128, CIIIλ1175, NV λ1240,
SiIV λ1400, et CIV λ1550 [21,22,23,24]. Les étoiles chaudes
développent des vents stellaires denses et rapides dus à la pression de radiation des photons UV sur les métaux de la photosphère stellaire. Il en résulte des profils de raies de type
P Cygni dans l’UV. Ces profils montrent une absorption fortement décalée vers le bleu (jusqu’à 3000 km/s) assortie d’une
émission du côté rouge. L’intensité et la largeur des profils sont
très sensibles à la température et gravité de surface et à l’abondance des métaux de l’étoile. Dans le spectre intégré d’une
population stellaire jeune, la forme des raies devient un excel-
252 C PHYSICS
IN
lent diagnostic des caractéristiques – âge, métallicité, fonction
de masse initiale et mode de formation – de la population stellaire présente.
Codes de synthèse évolutive
Sekiguchi & Anderson [25] sont les pionniers de la synthèse
ultraviolette pour l’étude du contenu stellaire de sursauts
éloignés, i.e. pour lesquels on ne peut résoudre les étoiles individuellement. Ils ont construit des spectres synthétiques UV
de diverses populations stellaires en additionnant des spectres
d’étoiles OB individuelles. Ils comparaient ainsi les largeurs
équivalentes des raies SiIV λ1400 et CIV λ1550 avec les
observations des galaxies. Mas-Hesse & Kunth [26] ont perfectionné cette technique en considérant l’évolution des étoiles en
fonction de l’âge de la population synthétique. Avec l’avènement du télescope spatial Hubble en 1990, une meilleure résolution spectrale et un meilleur signal pour des objets éloignés
permettent maintenant de bénéficier du plein potentiel du profil particulier des raies stellaires. Des bibliothèques spectrales
ultraviolettes (spectographe FOS) ont été assemblées pour les
étoiles de la Voie lactée et des Nuages de Magellan
(Robert et al. [27], Leitherer et al. [28,29], de Mello et al. [30]) et
ont été ajoutées au code de synthèse évolutive Starburst99
(Leitherer et al. [31-33], Varquez et al. [34]) et LavalSB
(Dionne [35], Dionne & Robert [36]). La synthèse dans l’UV
lointain a été entreprise en parallèle avec les missions
Copernicus et HUT (Gonzalez-Delgado et al. [18],
Robert et al. [37]).
LavalSB est une version parallèle du code Starburst99, qui a
été optimisée pour la synthèse UV à quatre métallicités différentes et pour tenir compte de la présence de systèmes
binaires massifs dans les populations stellaires. En résumé, ces
codes utilisent des tracés évolutifs (issus de modèles d’évolution stellaire) pour suivre les étoiles en fonction de l’âge et de
la métallicité de la population stellaire. Cette population est
d’abord définie par une FMI, un mode et un taux de formation
stellaire. La FMI, qui spécifie la distribution de masses des
étoiles dans un échantillon donné, est bien représentée par une
loi de puissance, ξ(M) % M−α; elle est caractérisée par une
pente α et des masses limites inférieure et supérieure (Mlo et
Mup). Le mode de formation stellaire peut être instantané
(toutes les étoiles apparaissent en même temps) ou continu (de
nouvelles étoiles s’ajoutent en même temps que les anciennes
évoluent). À chaque intervalle de temps (petit par rapport au
rythme d’évolution des étoiles), un spectre synthétique est
alors créé en additionnant les contributions des étoiles individuelles. Les caractéristiques stellaires (masse, température
effective, gravité de surface, abondance…) données par les
tracés évolutifs permettent, pour chaque étoile, d’assigner un
type spectral, de sélectionner un modèle d’atmosphère stellaire
et de calibrer en flux un spectre normalisé de la bibliothèque.
Finalement, la comparaison des raies synthétiques avec celles
observées donne des informations sur l’âge, la métallicité et la
FMI du starburst. La comparaison de la distribution d’énergie
UV globale avec le meilleur modèle permet d’estimer l’extinction intrinsèque causée par la poussière interstellaire et
LES GALAXIES À SURSAUTS ... (ROBERT)AA
d’obtenir une limite sur la masse stellaire impliquée. Plusieurs
observables peuvent aussi être prédits à partir du meilleur modèle obtenu. Par exemple, on peut calculer la force des raies
nébulaires (d’après le nombre d’étoiles massives présentes
et les photons ionisants qu’elles produisent) et la quantité
d’énergie, la masse et la composition chimique de la matière
retournée dans le milieu interstellaire. La technique de synthèse spectrale de l’UV a été appliquée depuis à un grand
échantillon de galaxies (Pellerin & Robert [19],
Gonzalez-Delgado
et
al. [39],
Chandar et al. [38],
Johnson et al. [40], etc.) et certains de ces résultats sont présentés dans la section suivante.
λ1400 se transforme en une absorption photosphérique large.
L’amplitude des profils diminue généralement si l’abondance
en métaux des étoiles est plus faible. Dans le domaine de l’UV
lointain, les raies CIII λ1175, PV λλ1118,1128 et
SiIV λλ1122,1128 développent similairement des profils
P Cygni avec la présence des étoiles chaudes. La raie
OIV λλ1032,1038 montre aussi un profil de type P Cygni dans
les vents stellaires [37], mais elle n’est pas aussi sensible aux
paramètres des étoiles et de la population stellaire. De plus,
lorsqu’on se déplace vers les plus petites longueurs d’onde,
l’absorption interstellaire (particulièrement due aux raies de
l’hydrogène moléculaire) devient très importante.
Les Figures 1 à 3 montrent un ensemble de spectres UV synthétiques normalisés, issus de modèles de sursauts instantanés
ayant diverses caractéristiques. Entre autres, on remarque que
la raie CIV λ1550 montre déjà un profil P Cygni pour une
population jeune de 1 Ma, ce profil devient cependant très
intense pour une population âgée de 3 Ma et si des étoiles massives sont présentes, i.e. si Mup $ 40 M et α $ 2.35. En
comparaison, la raie SiIV λ1400 montre un profil P Cygni
seulement si le sursaut a un âge près de 2-5 Ma, i.e. lorsque des
supergéantes O sont présentes. Lorsque les étoiles de type B
dominent la population stellaire, après 10 à 15 Ma, la raie SiIV
Pour un mode de formation stellaire continu, les raies stellaires
ne changent plus après qu’un équilibre ait été atteint entre
l’ajout de nouvelles étoiles et la mort des étoiles massives (vers
15-20 Ma). Des profils P Cygni avec une amplitude plus faible
que celle possible pour un sursaut instantané (due à la dilution
des raies par le continuum des vieilles étoiles qui s’additionnent avec le temps) sont alors prédits.
Fig. 1
Spectres synthétiques de l’ultraviolet proche pour des sursauts de métallicité et d’âge différents. Les modèles ont été
calculés avec LavalSB en utilisant la bibliothèque spectrale
construite avec des données du télescope Hubble. Les modèles considèrent un sursaut instantané avec une FMI ayant
une pente α = 2.35 et des masses entre 1 et 100 M . Deux
métallicités sont représentées : solaire (ligne pleine) et 1/10
solaire (traits pointillés). L’âge du sursaut est indiqué à côté
de chaque spectre. Les raies stellaires et interstellaires importantes sont identifiées au-dessus et au-dessous (respectivement) des spectres.
Caractérisation des populations stellaires
Dans leur ensemble, les diverses études de synthèses de galaxies permettent de tirer deux grandes conclusions : 1) la fonction de masse initiale semble universelle, avec une pente à la
Salpeter [41], i.e. α = 2.35, et une masse limite supérieure à
50 M , et 2) la durée de la formation stellaire ne dépasse pas
quelques millions d’années. On ne peut mettre de côté certains
Fig. 2
Spectres synthétiques de l’ultraviolet lointain pour des sursauts de métallicité et d’âge différents. Les modèles ont été
calculés comme pour la Figure 1 en utilisant ici la bibliothèque spectrale obtenue avec FUSE. Les raies interstellaires
de l’hydrogène moléculaire ne sont pas identifiées sur la figure, car elles sont trop nombreuses.
Fig. 3
Effet de la fonction de masse initiale sur les spectres synthétiques de l’ultraviolet proche. Les modèles ont été calculés
avec LavalSB en utilisant la bibliothèque spectrale construite
avec des données du télescope Hubble. Les modèles considèrent un sursaut instantané de 3 Ma, avec une métallicité
solaire et différentes pentes et masses limites supérieures de
la FMI. Les paramètres de la FMI sont indiqués au-dessus des
spectres.
une population stellaire de plus en plus dominée par des étoiles
de type B. Cependant, il faut aussi dire qu’une reproduction
acceptable des raies observées est possible dans le cas d’un sursaut continu, mais plus vieux, soit ~30 Ma avec Mup = 50 M .
Pour le meilleur modèle, instantané ou continu, l’extinction
interne E(B-V) doit être environ 0.45 (en adoptant la loi
générale d’extinction de Calzeti et al. [45] pour les starbursts)
afin de reproduire la pente théorique de la distribution spectrale
d’énergie de l’ultraviolet. Les modèles de synthèse de populations prédisent une pente du continuum qui varie peu avec les
caractéristiques (âge, métallicité et fonction de masse initiale)
d’un jeune sursaut [32]. Ainsi, la masse stellaire minimum qui
permet de reproduire le niveau de flux ultraviolet observé est
de l’ordre de 106 M . Si l’on suppose qu’un nombre de photons ultraviolets équivalent à ceux observés ionisent le milieu
environnant, le flux prédit dans la raie d’émission d’hydrogène
Hα est de l’ordre de 0.4x10-12 erg cm-2 s-1, ce qui est en bon
accord avec le flux observé [46].
biais et contraintes observationnelles lors de la sélection des
sursauts étudiés. Entre autres, un domaine de métallicité limité
et la qualité du signal voulue font en sorte que l’on choisit
d’abord les régions brillantes, donc de l’Univers local, mais
aussi déjà évoluées (i.e. sorties du cocon de matière où elles
naissent).
DE 30 DOR À NGC3690, EN PASSANT PAR NGC2363
La synthèse spectrale ultraviolette qui se basait sur les premières bibliothèques spectrales UV à haute résolution obtenue
avec le télescope Hubble, a d’abord été testée avec 30 Dor [42],
la région de formation d’étoile la plus intense du Grand Nuage
de Magellan. Ce sursaut étant proche, on y avait déjà identifié
individuellement les principales étoiles massives. Le satellite
IUE a été utilisé pour effectuer un balayage de R136, l’amas
central de 30 Dor, afin d’obtenir des spectres intégrés. La synthèse de ces données a redonné l’âge de 3 Ma déjà connu
pour ce sursaut et a bien reproduit les nombres d’étoiles O et B,
validant la technique pour des sursauts éloignés.
L’étude du spectre la région B2 [43] de la galaxie en interaction
NGC3690 observée avec le télescope Hubble, met bien en évidence la puissance et les limites de la technique de synthèse [44]. En adoptant une métallicité solaire pour cette région
et une pente standard pour la fonction de masse initiale, on peut
reproduire les profils des raies avec un sursaut instantané ayant
Mup = 50 M et un âge de 6.5 Ma (voir la Fig. 4). Dans ce cas,
on a peu de contraintes sur la fonction de masse initiale; pour
cet âge avancé, les étoiles massives ont déjà disparu si elles
étaient présentes initialement. Pour cette plage d’âges, la
dégénérescence de la solution en âge et en métallicité est
importante et on ne peut contraindre simultanément les deux
paramètres. Mais l’incertitude sur l’âge demeure faible (soit
1Ma) et la métallicité ne peut être beaucoup plus faible.
Comme le montre la Figure 5, si le sursaut est plus jeune que
6.5 Ma, les profils P Cygni vus dans les modèles sont trop forts
et s’il est plus vieux que 7 Ma, ils deviennent trop faibles pour
254 C PHYSICS
IN
Fig. 4
Synthèse spectrale ultraviolette de NGC3690. La figure
présente Arp299, un système en interaction dont fait partie
NGC3690. L’image du visible de Arp299 a été obtenue à
l’Observatoire du mont Mégantic par Daniel Devost. Le carré
sur cette image identifie approximativement la région de
NGC3690 qui a été observée dans l’UV avec l’instrument
FOS du télescope Hubble. La région B2 est identifiée, dans
l’ouverture de 1NN de l’instrument FOS, sur l’image UV. Le
spectre UV de B2 ainsi obtenu est représenté par une ligne
pleine foncée. Il tient compte du décalage cosmologique,
ainsi que du rougissement dû à la Voie lactée et à la galaxie
elle-même. Des spectres synthétiques sont présentés en traits
pointillés pour différents âges. Ils ont été calculés pour un
sursaut instantané à métallicité solaire, avec une FMI ayant
une pente α = 2.35 et des masses stellaires entre 1 et 80 M .
Lorsque le sursaut étudié est plus jeune que 5 Ma, il est plus
facile de distinguer les effets des paramètres de la FMI (pente
et masse limite supérieure) et de la métallicité. Par exemple, un
sursaut de 2-3 Ma ayant une faible métallicité de 0.1 Z reproduit bien le spectre observé de la région B de NGC2363 [47]. La
masse limite supérieure se doit de dépasser 60 M et la pente
de la FMI doit être égale ou inférieure à 2.35, i.e. des étoiles
LES GALAXIES À SURSAUTS ... (ROBERT)AA
massives sont nettement présentes. À ce jeune âge, le mode de
formation stellaire (instantané ou continu) ne peut être contraint, car l’effet de dilution par le continuum des vieilles
étoiles sur les raies des étoiles massives ne se remarque pas
encore.
UN MODE DE FORMATION STELLAIRE INSTANTANÉ
Le flux étudié dans le domaine de l’UV proche est souvent issu
de la combinaison de plusieurs générations d’étoiles, en amas
ou dans un fond diffus, dû à la dimension des ouvertures
disponibles. La synthèse spectrale qui ne considère qu’une
population unique est alors un peu simpliste ou a souvent pour
effet de rendre ambigu le mode de formation stellaire (instantané ou continu). Dans l’UV proche, il n’est pas rare en effet
d’identifier des raies spectrales associées à une population
dominée par des étoiles de type B et A (âgée d’environ 50 Ma)
en même temps que les signatures P Cygni des étoiles OB
(e.g. de Mello et al. [30]).
Les spectres obtenus avec le télescope FUSE, qui couvrent
l’UV lointain, malgré une ouverture très grande, semblent permettre d’isoler plus facilement un sursaut jeune et brillant et
d’éviter les générations sous-jacentes plus vieilles (car les
étoiles OB sont vraiment les principales contributrices du flux
de l’UV lointain). Dans leur étude, Pellerin et al. [19] ont considéré un échantillon de 24 galaxies observées avec FUSE,
couvrant un large domaine de métallicités. Pour la majorité de
ces objets, seul le mode instantané de formation stellaire arrive
à bien reproduire le spectre observé. Les galaxies qui font
exception sont mieux représentées en additionnant la contribution d’un deuxième sursaut instantané âgé d’environ 10 Ma,
plutôt qu’en utilisant un modèle continu. La Figure 5 est un
exemple de synthèse dans l’UV lointain de la galaxie
NGC7714 observée avec FUSE. Ce domaine de longueurs
d’onde favorise alors le mode instantané et suggère que la formation stellaire est de courte durée (moins de ~3 Ma).
Les systèmes binaires massifs
Certaines observations indiquent que la moitié des étoiles font
partie de systèmes multiples [48]. Dans le cas de systèmes
binaires massifs proches, un transfert de masse a lieu entre les
composantes, ce qui affecte par la suite leur évolution. Entre
autres, le gain de masse permet à une étoile initialement peu
massive de développer des vents stellaires, de devenir une
étoile de type Wolf-Rayet et de terminer sa vie en supernova.
Dionne & Robert [36] ont vérifié l’effet de ces systèmes sur la
synthèse des populations stellaires. Ils ont ajouté au code
LavalSB, des tracés évolutifs spécifiques pour les composantes
binaires. Ces tracés ont été adaptés des modèles d’évolution
pour les systèmes binaires massifs de de Loore et
Vanbeveren [49,50] pour les rendre compatibles avec les tracés
du groupe de Genève [51,52] déjà utilisés pour les étoiles sim-
Fig. 5
Synthèse spectrale ultraviolette de la galaxie NGC7714
observée avec FUSE. L’ouverture de 30NNx30NN a été centrée
sur le coeur de la galaxie. Le spectre tient compte du décalage
cosmologique, ainsi que du rougissement dû à la Voie lactée
et à la galaxie elle-même (E(B-V)int = 0.10). Le spectre synthétique (en rouge) a été calculé pour un sursaut instantané de
4.5 Ma, à métallicité solaire, avec une FMI ayant une pente
α = 2.35 et des masses stellaires entre 1 et 100 M .
ples. Dionne & Robert confirment ainsi, comme proposé par
d’autres études (e.g. Schaerer & Vacca [53]), un rapport WR/O
plus élevé à basse métallicité qui est alors en accord avec les
observations. Pour une fraction de systèmes binaires massifs de
l’ordre de 30%, les raies stellaires de l’UV ne sont pas modifiées de façon perceptible.
CONCLUSIONS
Notre capacité à décrire le contenu des galaxies est une étape
importante pour toutes les études visant à comprendre l’histoire
des galaxies et de l’Univers. La contribution de l’ultraviolet
dans ces travaux est unique en ce qui concerne les populations
jeunes, la fonction de masse initiale et le mode de formation
stellaire. En plus d’apporter des contraintes nous aidant à
établir des scénarios de l’évolution actuelle dans une galaxie,
les régions jeunes ont un impact considérable sur leur environnement et l’évolution qui va suivre. Le Canada en collaboration
avec les Indes lancera, vers la fin de l’année 2009, un nouvel
imageur à haute résolution spatiale pour l’ultraviolet, le
Ultraviolet Imaging Telescope (UVIT). L’avenir de la spectroscopie ultraviolette demeure cependant encore incertain. Le
développement récent d’instruments de type spectro-imageur à
grand champ dans le visible et l’infrarouge est aussi un apport
considérable pour étudier et localiser toutes les générations
d’étoiles, le gaz et la poussière, en relation avec les régions
jeunes de l’ultraviolet. Les années à venir seront des plus
prometteuses.
REMERCIEMENTS
C. Robert remercie le Conseil de recherches en sciences
naturelles et génie du Canada ainsi que le Fonds québécois de
la recherche sur la nature et les technologies pour le support
financier de ses étudiants et de ses travaux.
BIBLIOGRAPHIE
1.
2.
3.
T.M. Heckman, ASP Conf., 348, 467 (2006).
V. Buat, J. Iglesias-Paramo, & M. Seibert, et al., ApJ, 619, L51 (2005).
T.M. Heckman, ASP Conf., 148, 127 (1998).
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
19.
20.
21.
22.
23.
24.
25.
26.
27.
28.
29.
30.
31.
32.
33.
34.
35.
36.
37.
38.
39.
40.
41.
42.
43.
44.
45.
46.
47.
48.
49.
50.
51.
52.
53.
N.R. Walborn, IAU Symp., 148, 145 (1991).
C. Steidel, M. Giavalisco, M. Pettini, M. Dickinson, & K.L. Adelberger, ApJ, 462, L17 (1996).
A. Aloisi, M. Toisi, & L. Greggio, AJ, 118, 302 (1996).
D.W. Weedman, F.R. Feldman, V.A. Balzano, L.W. Ramsey, R.A. Sramek, & C.C. Wuu, ApJ, 248, 105 (1981).
N. Scoville, Ap&SS, 269, 367 (1999).
G.R. Meurer, T.M. Heckman, C. Leitherer, A. Kinney, C. Robert, & D.R. Garnett, AJ, 110, 2665 (1995).
A. Pellerin, M. Meyer, J. Harris, & D. Calzetti, ApJ, 658, L87 (2007).
R.C. Kennicutt Jr, K.A. Roettiger, W.C. Kell, J.R. van der Hulst, & E. Hummel, AJ, 93, 1011 (1987).
F. Combes, ASP Conf., 230, 213 (2001).
T.M. Heckman, RMxAC, 17, 47 (2003).
R.M. Gonzalez Delgado, Ap&SS, 303, 85 (2006).
A.L. Kinney, R.C. Bohlin, D. Calzetti, N. Panagia, & R.F. Wyse, ApJS, 86, 5 (1993).
B.C. Whitmore, F. Schweizer, C. Leitherer, K. Borne, & C. Robert, AJ, 106, 1354 (1993).
C. Leitherer, W.D. Vacca, P.S. Conti, A.V. Filippenko, C. Robert, W.L.W. Sargent, ApJ, 465, 717 (1996).
R.M. Gonzalez-Delgado, C. Leitherer, & T. Heckman, ApJ, 489, 601 (1997).
A. Pellerin, & C. Robert, MNRAS, 381, 228 (2007).
L. Bianchi, D.A. Thilker, D. Burgarella et al., ApJ, 619, L71 (2005).
N.R. Walborn, J.W. Parker, & J.S. Nichols, NASA Ref. Pub., 1363 (1995).
N.R. Walborn, J. Nichols-Bohlin, & R.J. Panek, NASA Ref. Pub., 1155 (1995).
N.R. Walborn, A.W. Fullerton, P.A. Crowther, L. Bianchi, J.B. Hutchings, A. Pellerin, G. Sonneborn, & A.J. Willis, ApJS, 141, 443
(2002).
A. Pellerin, A.W. Fullerton, C. Robert, J.C. Howk, J.B. Hitchings, N.R. Walborn, L. Bianchi, P.A. Crowther, & G. Sonneborn, ApJS,
143, 159 (2002).
K. Sekiguchi, & K.S. Anderson, A&A, 94, 644 (1987).
J.M. Mas-Hesse, & D. Kunth, A&AS, 88, 399 (1991).
C. Robert, C. Leitherer, & T.M. Heckman, ApJ, 418, 749 (1993).
C. Leitherer, C. Robert, & T.M. Heckman, ApJS, 99, 173 (1993).
C. Leitherer, J.R.S. Leao, T.M. Heckman, D.J. Lennon, M. Pettini, & C. Robert, ApJ, 550, 724 (2001).
D.F. de Mello, C. Leitherer, & T.M. Heckman, ApJ, 530, 251 (2000).
C. Leitherer, C. Robert, & L. Drissen, ApJ, 401, 596 (1992).
C. Leitherer, & T.M. Heckman, ApJS, 96, 9 (1995).
C. Leitherer, D. Schaerer, J.D. Goldader et al., ApJS, 123, 3 (1999).
G.A. Vazquez, C. Leitherer, D. Schaerer, G. Meynet, & A. Maeder, ApJ, 664, 995 (2007).
D. Dionne, IAU Symp., 193, 596 (1999).
D. Dionne, & C. Robert, ApJ, 641, 252 (2006).
C. Robert, A. Pellerin, A. Aloisi, C. Leitherer, C. Hoopes, & T.M. Heckman, ApJS, 144, 21 (2003).
R. Chandar, C. Leitherer, C.A. Tremonti, D. Calzetti, A. Aloisi, G. Meurer, & D. de Mello, ApJ, 628, 210 (2005).
R.M. Gonzalez-Delgado, C. Leitherer, G. Stasinska, & T.M. Heckman, ApJ, 580, 824 (2002).
K.E. Johnson, C. Leitherer, W. Vacca, & P.S. Conti, ApJ, 120, 1273 (2000).
E.E. Salpeter, ApJ, 121, 161 (1955).
W.D. Vacca, C. Robert, C. Leitherer, & P.S. Conti, ApJ, 444, 647 (1995).
C.G. Wynn-Williams, K.W. Hodapp, R.D. Joseph et al., ApJ, 377, 426 (1991).
C. Robert, IAU Symp., 193, 616 (1999).
D. Calzetti, A.L. Kinney, & T. Storchi-Bergmann, ApJ, 458, 132 (1995).
R.D. Gehrz, R.A. Sramek, & D.W. Weedman, ApJ, 267, 551 (1983).
L. Drissen, J.R. Roy, C. Robert, D. Devost, & R. Doyon, AJ, 119, 688 (2000).
B.D. Mason, D.R. Gies, W.I. Hartkopf et al., Ap, 115, 821 (1998).
C. de Loore, & D. Vanbeveren, A&AS, 103, 67 (1994).
C. de Loore, & D. Vanbeveren, A&AS, 105, 21 (1994).
G. Schaller, D. Schaerer, G. Meynet, & A. Maeder, A&AS, 96, 269 (1992).
G. Meynet, A. Maeder, G. Schaller, D. Schaerer, & C. Charbonnel, A&AS, 103, 97 (1994).
D. Schaerer, & W.D. Vacca, ApJ, 497, 618 (1998).
256 C PHYSICS
IN
ARTICLE DE FOND
BRINGING CHEMISTS AND PHYSICISTS TOGETHER:
THE LEGACY OF THE ONTARIO PHOTONICS CONSORTIUM
AT THE UNIVERSITY OF WESTERN ONTARIO
BY
ROBERT H. LIPSON
L
oosely speaking, photonics is the science of generating, manipulating and detecting photons and,
as such, may be considered the purview of the
physics community. However, inherent in photonics research is the need for new and novel materials.
Consequently, many of the most exciting advances in photonics are coupled strongly to advances in nanomaterial
syntheses and applications. Enter the chemists! Of course
this is a simplistic division of labour which also ignores
the important contributions coming out of many engineering and biomedical departments.
The photonics community in Canada has become highly
organized in the area of photonics. The Canadian
Photonics Consortium is the representative voice of the
entire Canadian Photonics community. Membership
includes large and small companies, large and small academic institutions and consortia, and government laboratories and agencies at both the federal and provincial levels. Their stated vision (http://www.photonics.ca/community.html) is “to establish Canada as the place for business
success in Optics and Photonics”.
At the federal level, a National Centre of Excellence: the
Canadian Institute for Photonics Innovation (CIPI,
http://www.cipi.ulaval.ca/) was established in 1999, and
currently funds more than 100 researchers (many in chemistry and physics) and 250 students at 21 universities in
projects involving photonics, biophotonics, and information and telecommunications.
The Ontario government has been a particularly strong
supporter of photonics through their Ontario Centres of
Excellence program (OCE, http://www.oce-ontario.
org/Pages/ COEPhotonics.aspx?COE=PH). The program
mandate is to facilitate economic growth through support
for industrially relevant research and development, and to
open new market opportunities and the commercialization
of leading edge discoveries. The Ontario Photonics
Consortium (OPC) was funded by the older provincial
Ontario Research and Development Fund in 2000. This
initiative brought together chemists, physicists and engineers from the University of Western Ontario, McMaster
University, the University of Waterloo, the University of
Toronto, and the University of Ottawa. Historically, the
consortium was established as a result of a $45M investment by the Ontario Government for a proposal synthesized from three earlier separate requests; the first dealing
with fundamental science in the area of photonic band gap
materials (UWO, lead PI Ian Mitchell); the second with
photonic devices (McMaster, lead PI Peter Mascher) and
the third also from McMaster (lead PI, Wei-Ping Huang)
dealing with large systems. While the objectives of the
three proposals were quite distinct, it was also recognized
that innovations and breakthroughs in any one of the three
areas would positively impact the others, and hence, the
merger. The last five years have clearly demonstrated that
the combined group of researchers benefitted not only by
having a wider network to interact with, but from also having access to relatively unique facilities at the different
institutions.
The majority of the photonics and nano activities taking
place at Western began because of the opportunities that
arose when the CFI-funded Nanofabrication Laboratory
(http://www.uwo.ca/fab/) opened in September 2004.
This class 100 cleanroom facility houses a suite of instrumentation including SEM imaging, FIB lithography, silicon DRIE, ion beam implantation and analyses, and TEM
specimen preparation. Many of the applications initiated
through OPC funding involved lithographic patterning of
novel materials for plasmonic, photonic band gap (PBG)
materials, and biophotonics applications.
R.H. Lipson
<[email protected]>,
Dept. of Chemistry,
University of Western
Ontario, London, ON
N6A 5B7
PLASMONICS
SUMMARY
Interest and activities in photonics worldwide have exploded over the last decade
because of its commercial potential, and
because of the synergy it inspires between
the chemistry and physics communities.
Plasmons are electron density waves created when optical
light hits the surface of a metal. Plasmonics can be considered the study and application of the transfer of energy
between the light and electrons. Ian Mitchell (Physics)
and Kim Baines (Chemistry) with an undergraduate student Michelle Watroba initiated a collaboration to test
optical scattering theory by examining the light scattering
BRINGING CHEMISTS AND PHYSICISTS TOGETHER ... (LIPSON)
Fig. 1
a) Scanning electron micrograph of a single gold-coated silica nanoparticle; b) Controlled deformation of a nano-sphere
by ion-beam exposure.
properties of 2-D lattices formed by metallodielectric nanoparticles arranged into an array on a lithographically-defined
PMMA substrate. This work takes advantage of Baines’ expertise in organometallic chemistry to make the nanoparticles
(Fig. 1a) and Mitchell’s extensive background in ion-beam
physics. The light scattering particles consisted of spherical
silica cores of submicron diameter (nanospheres) coated with a
gold shell to a selected thickness ranging from tens through
hundreds of nanometers. Experiments were also carried out
with Senior Research Scientists Todd Simpson in the Nanofab
to test the sensitivity of scattering to alteration of the shape of
the array by driving the nanosphere into an ellipsoidal shape by
ion-beam exposure (Fig. 1b). An array of spheres is shown in
Fig. 2.
In related work,
p h y s i c i s t
Silvia Mittler working with chemist
Zhifeng Ding have
characterized
the
electrochemistry of
self-assembled monolayers (SAMs) of
m o n o m e r i c
calix[4]arenes and
Fig. 2 Scanning electron micrograph of h e t e r o d i m e r i c
a 2-D array of gold-coated silica calyx[4]arenes capnanoparticles.
sules filled with ferrocenium
(shown
schematically in Figure 3) on Au surfaces for data storage purposes [1].
This molecular guest host systems can be filled with a variety
of guest molecules. OMCVD (organometallic chemical
vapour deposition) grown gold nanoparticles coated with these
calix[4]arene heterodimer capsules leads to distinct surface
plamon resonances whose spectral position depends on the
dielectric constant of the guest molecule. Mittler and her group
in cooperation with Patrick Ronney and Chitra Rangan from
the Department of Physics at the University of Windsor could
show experimentally and theoretically how the surface plasmon resonance shifts by systematically varying the dielectric
function of a monolayer on a gold nanoparticle with fixed
thickness.
258 C PHYSICS
IN
Fig. 3
Structure of calix[4]arene heterodimers on gold.
The Mittler-Rangan team have also examined how the surface
plasmon spectrum depends on the proximity of the nanoparticles with respect to each other, in both air and water environments. The wavelengths of the spectral features are strong
functions of the distance dependences of the electromagnetic
dipolar and quadrupolar interactions between the particles. As
shown in Figure 4 the experimental spectra are in excellent
agreement with the simulations. These results are expected to
valuable for optimizing sensor applications.
Fig. 4
Calculated extinction spectra for nanoparticle pairs with
varying separations. Particles are (top left) uncoated in air,
(top right) coated in air, (bottom left) uncoated in water and
(bottom right) coated in water. Coatings have a refractive
index of 1.45 and a thickness of 1.75 nm. The particle radius
is 7 nm. The separation is measured as the distance between
surfaces, which for coated particles corresponds to the coating surface,not the nanoparticle surface.
More recently, François Lagugné-Labarthet (Chemistry) was
recruited to Western to develop novel surface spectroscopies.
One technique that has already begun to show great promise is
Tip-Enhanced and Surface–Enhanced Raman spectroscopy
(TERS and SERS, respectively) for the detection of biomolecules on surfaces [2]. Part of his overall strategy involves using
the Nanofab to fabricate SERS substrates with reproducible
behaviours.
Lagugné-Labarthet and graduate student
Betty Galarreta have made well-controlled lithographic patterns of noble metals (Au) on glass as shown in Fig. 5a-b. The
BRINGING CHEMISTS AND PHYSICISTS TOGETHER ... (LIPSON)AA
inter-structure gap between the features (between 30 and
50 nm) is a key parameter for optimizing the SERS enhancement via the surface plasmon resonance of the metal.
Depending on the sizes and gaps of the structure, the plasmon
frequency can be finely tuned as shown in figure 5c. Figure 5d
shows an example of a SERS spectrum of a monolayer of
guanosine triphosphate (GTP) measured under a confocal
microscope. The spectra are significantly enhanced when
deposited on their platform while it appears very weak on a flat
gold surface. The vibrational information contained in the
SERS spectrum can provide invaluable information about their
insertion into the biological membranes, their structural conformations, or their interactions with surrounding molecules,
and are important for understanding many fundamental bioprocesses.
enhance its ability to serve as a platform for cell adhesion in
microfluidic devices [3]. The process involves depositing a thin
layer of aluminum onto the polymer in the presence of an Ar
plasma. The PDMS surface can then be patterned by depositing metal through a stencil mask thereby allowing cells to
adhere to specific locations on the surface (Figure 6). Such
arrays allow the study of cell-cell interactions, cell motility,
and cellular responses to various spatial and geometric perturbations.
Zhifeng Ding is developing improved and integrated tools to
investigate single live cells and semiconductor nanostructures
to provide insight into
the
relationships
between structure and
property and or function [4-7]. One important approach is the
use of Scanning
Electrochemical
Microscopy (SCEM)
which involves measuring the current of
species contained in
the solution gap
between a tip and the Fig. 7 SECM images of two COS 7 live
cells
substrate. SECM is
useful in a wide range
of applications, including imaging of biological molecules.
For example, Figure 7 shows an SECM image of two COS 7
live cells from which information about their metabolism can
be derived.
NOVEL MATERIALS
Fig. 5
Examples of plasmonic devices made using the e-beam
lithography technique. a) the sharp structures of the nanosnowflakes have gaps in the 50-100 nm range. b) The triangles have a side of 400 nm and a 20 nm gold thickness. c) The
Plasmon frequency can be tuned very accurately depending
on the size of the triangles and associated gaps. d) Raman
spectra of a monolayer of GTP deposited on flat gold and on
SERS platforms. The laser power is similar in both experiments.
BIOPHOTONICS
Peter Norton
(Chemistry),
Nils Petersen (NINT)
and graduate students
Jessica McLachlan,
Natasha Patrito,
and post-doc
Claire McCague have
developed a method
for the modification
of the surface of
poly(dimethylsiloxane), PDMS, to
Fig. 6
Optical micrograph of patterned
C2C12 cells on modified PDMS.
Novel materials, fabrication techniques, and applications of
photonic crystals (PCs) are core areas of research at Western.
Rob Lipson and Kim Baines joined forces with co-supervised
student Yun Yang to examine the possibility of fabricating PCs
by optical lithography
in photoresists made
from Si- or Ge-containing polymers [8].
PCs have periodic
structures
which
localize light and prohibit a certain range
of wavelengths from
propagating within
the material. The PBG
structures are made Fig. 8 SEM image of a germanium thin
from two substances
film (poly(p-methoxyphenylwith significantly difmethylgermane) patterned by 2beam interference lithography.
ferent
refraction
indices, n. In a similar
manner to semiconductors which exhibit electronic band structure due to their periodic atomic spatial arrangement, the periodic variation of the index of refraction in a PC produces a
band structure for photons, with well-defined energy-momen-
tum levels. One condition that is required to open a band gap
is that the periodic index contrast of the crystal must be large
(Δn~ 2). This can not usually be achieved using commercially
available carbon-based photoresists. Instead patterned carbonbased resists are usually used as masks for subsequent etching
into high index substrates such as Si. The Baines group was
able to synthesize a series of photosensitive polymers having Si
and Ge backbones. As shown in Fig. 8, these polymers could
be patterned by optical lithography. The indices of refraction
of Si- and Ge-based films are sufficiently large that their use
photoresists can in principle lead to photonic band gap structures in a single fabrication step.
Martin Zinke-Allmang working with student Kenneth
Kar Ho Wong, and Engineering Professor W.K. Wan are examining the elastic properties of non-woven fibres with nanometer diameters using
high resolution scanning electron microscope (SEM) and Xray
photoelectron
spectroscopy
(XPS) [9]. These polymers are becoming
the biomedical materials of choice in
many restorative and
Fig. 9 A SEM image of a PVA fibres regenerative medical
procedures because
mat
their physical properties, such as porosity and mechanical strength, can be tailored
to suit specific applications. In recent experiments, an electrospinning process was used to produce non-woven polymer
mats composed of fibres with diameters between 50 nm and
500 nm. Figure 9 shows a mat of poly(vinyl alcohol) (PVA)
formed in this way. Elastic moduli were found using the
clamped beam model to fit the deflection values along the suspended fibre after some accurate measurements of the geometry and diameter of the fibre were established.
Lipson and Cheng Lu
have studied the synthesis and optical
properties of thin
films of β-barium
borate (β-BaB2O4;
β-BBO). They have
made high quality
thin films of β-BBO
which are amenable
to contact lithography
(Figure 10) [10]. The
films are produced by
spin coating metalloorganic solutions with
a poly(vinyl pyrrolidone) (PVP) additive,
followed by O2 plas-
260 C PHYSICS
IN
Fig. 10 SEM images of a) the Si mask
used for contact lithography and
b) the patterned β-BBO film.
ma treatment and thermal baking. In addition, the BBO thin
films could be reoriented by seeding the precursor gels with an
organic molecule prior to thermal treatment. Using either
approach, the films exhibit more efficient second harmonic
generation than those made by literature methods.
In different experiments, new routes to thin films of solid state
VO2 have also been developed using sol-gel methods. By precisely controlling the
processing
conditions, (baking temperature, ambient gas in
the oven, baking time,
solvent etc.), VO2
films can be synthesized that are highly
resistant to oxidation
for long periods of
time. Furthermore,
by varying the processing conditions,
the morphology of the
resultant VO2 films Fig. 11 Raman image of nanowires of
can be controlled to
vanadium oxide obtained by
produce of nano-belts,
detecting specific Raman modes
of the material
nano-ribbons or nanowires made of V2O5.
These materials have in part been characterized using Raman
imaging in collaboration with the Ding group (Figure 11). VO2
undergoes a phase change from semiconductor to metal near
70oC on the picosecond time scale. The films produced at
Western are being examined by OPC member Dwayne Miller
at the University of Toronto using femtosecond electron diffraction [11] to better understand this remarkable transition.
Zhifeng Ding’s group,
in collaboration with
T . - K . S h a m
(Chemistry)
and
Xueliang Sun
( M a t e r i a l s
Engineering)
has
found an electrochemical avenue to
prepare strong blue
l u m i n e s c e n t
nanocrystals (NCs)
from multiwalled carFig. 12 UV-visible absorption and photobon
nanotubes
luminescence spectra of carbon
( M W C N T s )
NCs in aqueous solution. Inset is
The
(Fig. 12) [12].
the solution illuminated by an
search for a good carUV lamp.
bon emitter is a challenging enterprise because neither of the two bulk carbon
allotropes, graphite and diamond, give strong luminescence.
The new carbon NCs prepared at Western are very attractive
due to their promised applications in optoelectronic devices,
biology labelling, and biomedicine.
BRINGING CHEMISTS AND PHYSICISTS TOGETHER ... (LIPSON)AA
PHOTONIC CRYSTALS
Lipson, Mitchell and graduate student Cheng Lu have developed novel optical lithography techniques that are expected to
ultimately be used for fabricating PCs. In one approach nearfield Diffraction Element Assisted Lithography or DEAL was
devised to fabricate two-dimensional lattice patterns in a photoresist [13]. Specifically, a diffraction element was used to prepattern the coherent output of a laser prior to its capture in a
photoresist. The pattern symmetry and spacing can be readily
modified with the same experimental arrangement since the
near-field diffraction pattern strongly depends on the nature of
the diffractive element and the distance between the element
and the photoresist. The patterns that are formed can serve as
masks for patterning high index materials to create photonic
band gap materials. Alternatively, they have the potential to
behave as two-dimensional photonic band gap arrays provided
the polymer used exhibits a large enough index contrast.
In a second approach, a Babinet-Soleil compensator was inserted into the path of one of the three beams used for noncoplanar
beam interference lithography [14]. This birefringent element
could change the phase of the beam so that either a positive
two-dimensional pattern or an inverse-like structure is generated in a photoresist without disturbing the mechanical geometry
of the setup. As shown in Figure 13 large defect free sample
areas (>1cm2) with sub-micron periodic patterns with different
morphologies could obtained by simply “dialing” up a specific
phase difference for one of interfering beams.
Among the diverse photonic crystal (PC) applications, PC sensors have drawn much attention because of their high sensitivity and compact structure [15]. Jayshri Sabarinathan
(Engineering) has developed a PC waveguide based pressure
sensor which has many applications in MEMS and microfluidic applications. Sensing is performed by measuring the transmission variation through the PC waveguide due to the changes
in the refractive index of the region surrounding the PC. When
pressure is exerted on the waveguide it mechanically deforms
the waveguide and alters the transmission characteristics of the
waveguide. The changes in light intensity due to the relative
displacement of the PC waveguide with respect to substrate can
be correlated to the fluid pressure.
The device shown in
Fig. 14 consists of an
air bridge PC waveguide with a triangular air hole lattice
coupled with conventional dielectric waveguides on input and
output side, designed
on a silicon-on-insulator (SOI) wafer.
Simulations
show Fig. 4 SEM micrograph of a Photonic
Crystal air-bridge waveguide for
nearly 72% and 0%
pressure sensor applications
transmission when the
distance between the
PC waveguide and the substrate was 600nm and 0nm respectively. Large intensity variations with small displacements
were achieved when the PC waveguide was between 300 nm to
200 nm.
CONCLUSIONS
Fig. 13 a) SEM image of a pattern formed in a SU-8 photoresist
when all three beams used for IL had the same phase, (φ1, φ2,
φ3) = (0, 0, 0). The marker indicates a 3 μm scale; b) SEM
image of a pattern formed in a SU-8 photoresist when the
phase of the third beam was π/2 different from the other two,
(φ1, φ2, φ3) = (0, 0, π/2). The bar marker indicates a 3 μm
scale.
The examples above constitute only a very small subset of
those that continue to develop even though the formal activities
of the OPC have concluded. They show that photonics studies
are a platform for strong collaborations between chemists,
physicists and beyond. The work has strong fundamental and
applied relevance, and therefore, it is expected that the resultant partnerships are more long term than short. In this regards,
the future of photonics and the synergy it generates between
different communities is bright indeed.
REFERENCES
1.
2.
3.
4.
5.
6.
7.
8.
9.
S. Xu, G. Podoprygorina, V. Boehmer, Z. Ding, P. Rooney, C. Rangan, S. Mittler, Organic & Biomolecular Chemistry 5 558-568,
2007.
V. Guieu, F. LagugnépLabrathet, L. Servant, D. Tulaga, and N. Sojic, Small, 4, 96-99 2008.
Natasha Patrito, Claire McCague, Peter R. Norton, Nils O. Petersen, Langmuir 23, 715-719, 2007.
R. Zhu, Z. Ding, Can. J. Chem. 83 1779-1791, 2005.
X. Zhao, N.O. Petersen, Z. Ding, Can. J. Chem. 85 175-183, 2007.
P.M. Diakowski, Z. Ding, Phys. Chem. Chem. Phys. 9 5966 – 5974. 2007
R. Zhu, Z. Qin, J.J. Noel, D.W. Shoesmith, Z. Ding, Anal. Chem. 80 1437-1447, 2008.
Y. Yang, M.Sc. Thesis, The University of Western Ontario, 2008
K.H. Wong, M. Zinke-Allmang, W.K. Wan, J.Z. Zhang, P. Hu, Nuclear Instruments & Methods in Physics Research, Section B: Beam
Interactions with Materials and Atoms 243, 63-74, 2006.
10.
11.
12.
13.
14.
15.
C. Lu, S.S. Dimov, and R.H. Lipson, Chem. Mater. 19, 5018-5022, 2007
B.J. Siwick, J.R. Dwyer, R.E. Jordan, R.J.D. Miller, Science, 302, 1382-1385.
J. Zhou, C. Booker, R. Li, X. Zhou, T.-K. Sham, X. Sun, Z. Ding, J. Am. Chem. Soc. 129 744-745, 2007.
C. Lu, X.K. Hu, I.V. Mitchell and R.H. Lipson, Appl. Phys. Lett. 86, 193110-1 – 193110-3, 2005.
C. Lu, X.K. Hu, S.S. Dimov and R.H. Lipson, App. Opt. 46, 7202-7207, 2007.
S. Mittal and J. Sabarinathan, Proceedings of the SPIE, 5971, 59711J, 2005.
262 C PHYSICS
IN
IN MEMORIAM
MARTIN WESLEY JOHNS - (1913-2008)
At McMaster he was Chair of the Department of Physics for
a total of twelve years, including the period when universities expanded rapidly in the 1960’s, and his foresight was
important in selecting about twenty new faculty members to
form a strong and harmonious department which proved to
be productive for the decades since. As his time for normal
mandatory retirement approached he was asked to continue at McMaster and serve as Coordinator of Part Time
Degree Studies. During his four years in this position (19771981) he expanded the size and popularity of that department considerably.
His research in the field of Nuclear Structure Physics, using
experimental techniques of beta and gamma spectroscopy,
is well respected internationally. He supervised the construction of several spectrometers and was a prime mover
in establishing the Tandem Van de Graaff Accelerator
Laboratory at McMaster. He co-authored about one hundred papers in scientific journals, received honourary D.Sc.
degrees from Brandon University and McMaster, and in
1958 was admitted to the Royal Society of Canada.
In 1960, during a Sabbatical leave at Oxford, he was sent to
Pakistan for nearly three months as a technical advisor for
the Canadian Department of External Affairs, to assess
whether Pakistani nuclear scientists and researchers possessed the expertise necessary to handle the nuclear reactor that Pakistan was requesting from Canada as part of the
Colombo plan.
Martin also had a very active role in Community Services
outside the University. He was named “Distinguished
Citizen of the Year for 1978” for Hamilton, Ontario, in part
due to his service to the United Way and the Family Service
Agency. He was an active supporter of the Hamilton United
Way, serving as both the Chair of its Allocations Committee
and as President of its Board for the Hamilton-Burlington
area. For many years he was heavily involved with the
Family Service movement, having served as President of
Family Services of Hamilton Wentworth, as well as
President of the Family Service organization at the provincial (Ontario) and federal levels.
He actively supported Westdale United Church in Hamilton
for over sixty years, and served the United Church of
Canada in several positions up to the national level. In 1999
Westdale United Church named its church hall “The Martin
Johns Hall” in view of his years of selfless work and generosity.
One project of which he was most proud in recent years was
the Campus Ministries Council he established and supported generously at McMaster University. It is an interfaith
chaplaincy which serves all Christian denominations and
non-Christian communities as well. The interfaith and interracial work of this group is an example of Martin’s influence
and efforts, which continued more than 25 years after his
official retirement.
Martin was always very generous with his time, talents, and
resources, and had a well-deserved reputation for treating
people fairly. One notable skill was his ability to understand
and deal with people. While chairing a committee or department he would usually manage to achieve consensus on
difficult issues, avoiding political maneuvering or narrow
votes that would leave some people dissatisfied. This skill
was also used widely in his everyday life. Students would
often consult him for advice on personal issues as well as
academic matters. He was held in great respect by his
grandchildren, who confided in him and communicated regularly by e-mail, discussing personal problems and asking
for advice.
His well-rounded outlook and range of interests is shown by
some of his other activities. For many years he sang in the
Bach-Elgar choir, and held season’s tickets for the Hamilton
Philharmonic Orchestra, Opera Hamilton, and the Hamilton
Tiger Cats Football games. At his Lake Boshkung summer
cottage in the Haliburton region he enjoyed carpentry and
sailing, and in his sixties took up windsurfing. Also, at that
age he could more than hold his own on the squash court
against students who were forty years younger! In later
years he learned to use a computer and wrote three autobiographical books. The first of these, “Bamboo Sprouts and
Maple Buds”, provides many interesting insights from his
early years in China. (It has recently been reprinted in hardcover by his granddaughters Sarah Turner and Alison
Crump, and is available for purchase online at
www.lulu.com/content/4059218/.)
Martin was part of a family that was very active
in the academic and professional communities
of Canada. After their return from China his
father, Alfred Johns, held Faculty positions in
the Departments of Mathematics at Brandon
College (1927-1931) and McMaster University
(1931-1952). Martin had three younger brothers and one sister. His brother Harold was well
known to the scientific community for developing the first Cobalt-60 radiation therapy unit (at
Saskatoon), and for his many years of work in
Medical Physics at Toronto. Paul had a career
in meteorology, and Edward was an orthodontist. Ruth was trained as a social worker and
spent many years working in that field. Martin
is survived by his sister Ruth, brother Paul,
daughter Beth, son Ken, nine grandchildren
and two great-grandchildren.
Dennis Burke
Professor, McMaster University, Retired
IN MEMORIAM
Martin Wesley Johns, Professor
Emeritus of Physics at McMaster
University,
passed
away
on
September 18, 2008, in his 96th
year. He was well known for his long
lifetime of unselfish service to the
scientific, university, church, and
social services communities of
Canada. Martin was born in China in
1913 of Canadian missionary parents, and was 12 years old when his
family returned permanently to
Canada. He completed his B.A. (1932) and M.A. (1934) at
McMaster University, and his Ph.D. (1938) at the University
of Toronto. His academic and scientific career then included
nine years as Professor of Physics at Brandon College
(1937-1946), a year at the Chalk River laboratories (19461947), and 34 years at McMaster University (1947-1981).
Dennis Burke
<dgp@physics.
mcmaster.ca>
is a Professor
Emeritus at the
Department of
Physics and
Astronomy,
McMaster
University, 1280
Main St. W.,
Hamilton, Ontario,
L8S 4M1
CAP OFFICE
CAP NEWS / INFORMATIONS DE L’ACP
3RD IUPAP INTERNATIONAL CONFERENCE ON WOMEN IN PHYSICS
(ICWIP-2008)
The 3rd IUPAP International Conference on Women in Physics (ICWIP 2008) was held during October 7-10, 2008 in Seoul,
Korea. This meeting takes place every three years and is organized under the auspices of the International Union of Pure and
Applied Physics (IUPAP) and the Korean Physical Society (KPS). The conference was dedicated to the presentation and discussion of the latest developments and ideas regarding the status of woman physicists in societies around the globe. Besides
attending plenary talks from many different research fields, the attendants had the opportunity to participate in workshops
on i) Individual professional development, ii) Attracting girls to physics and blocking the leaky pipeline, iii) Assessing and
improving the climate for women, iv) Successful proposals and project leadership, and v) Actions for Women in Physics
(WIP) working groups.
A press release issued by the ICWIP following the conference appears below. A copy of the paper and poster presented by
the Canadian delegation follow.
CANADIAN DELEGATION
BACK FROM INTERNATIONAL CONFERENCE
ON WOMEN IN PHYSICS
(ICWIP2008) IN KOREA
Roby Austin (St. Mary’s
University), Sampa Bhadra
Women, men, institutions,
(York University), Janis
and governments need to
McKenna (University of
work together to encourage,
British Columbia), Adriana
educate, recruit, retain,
Predoi-Cross (University of
advance, and promote more
Lethbridge), Michael Steinitz
girls and women in physics
(St. Francis Xavier Univerand other science and techsity), and Li-Hong Xu (Uninology professions. To that
versity of New Brunswick)
end, the conference particirecently returned from Seoul, Canadian delegation to ICWIP2008 (from left to right): Janis pants unanimously approved
Korea, where they were part McKenna (UBC), Roby Austin (St. Mary’s U.), Adriana a resolution presented at the
of over 330 scientists, from Predoi-Cross (U. Lethbridge), Li-Hong Xu (UNB), Sampa 26th General Assembly
nearly 70 countries from all Bhadra (York U.), and Michael Steinitz (St. F-X)
International Union of Pure
corners of the world, who
and Applied Physics in
took part in the Third IUPAP (International Union for Pure
Tsukuba, Japan on 15 October 2008.
and Applied Physics) International Conference on Women
Dr. Youngah Park, a physicist who chairs the conference
in Physics (ICWIP2008). Delegates came from African,
organizing committee, was recently elected to the Korean
Asian, European, Latin American, North American, and
National Assembly from her district. She told the assemisland nations.
bled participants, “I believe the positive effect of
ICWIP2008 will go beyond the physics community and
The meeting, held on October 7th to 10th, was dedicated to
will have a strong effect on women leaders in all fields of
celebrating the physics achievements of women throughout
science and technology”.
the world, networking toward new international collaborations, gaining skills for career success, and aiding the forThe First International Conference on Women in Physics
mation of active regional working groups to advance
was held in Paris in 2002. The Second conference was hostwomen in physics. Each country presented information
ed by Rio de Janeiro in 2005. Since the first conference
about its statistics and its activities to increase women's parmost countries have made some progress in attracting girls
ticipation.
to physics, increasing the proportion of physics degrees
awarded to women, and promoting the career development
Worldwide, fewer than 15% of physicists are women. More
of women physicists. However, the proportion of physicists
than 80% of the conference attendees were women. It was
who are women is well below 20% in nearly all countries clear that the scarcity of women in physics, especially in
too few to have maximum benefit for society.
leadership positions, is a problem for many countries.
CAP OFFICE
264 C PHYSICS
Nations cannot benefit fully
from women's ideas and
approaches to improve their
economic competitiveness, or
solve difficult problems, such
as energy, health, and global
sustainability.
IN
ENSEIGNEMENT
WOMEN PHYSICISTS
IN
CANADA
(PAPER PRESENTED BY THE CANADIAN DELEGATION TO
CONFERENCE ON WOMEN IN PHYSICS - ICWIP2008)
BY
THE
3RD IUPAP INTERNATIONAL
ADRIANA PREDOI-CROSS , ROBY AUSTIN, SAMPA BHADRA, JANIS MCKENNA, LI-HONG
XU AND MICHAEL STEINITZ
T
here is strong evidence that at the national level,
over the past decade the overall climate for
women physicists both in academia and industry
has improved. Organizations such as the
Canadian Association of University Teachers (CAUT) are
actively making efforts to minimize the current socioeconomic and professional gaps between women and men.
CAUT is also trying to ensure that women in academia at
all professional levels, are offered equal opportunities
with their male colleagues.
At the institutional level, in recent years several Canadian
Physics Departments have conducted an external critical
assessment of the climate and environment for women in
their physics departments. As a consequence, a friendly,
open, invigorating, welcoming climate towards women
colleagues was established and maintained. Some universities have made great strides and have 3 or 4 female faculty members. Unfortunately, across the country we still
have numerous departments that have no women faculty
members or have hired the first woman faculty member
after over 25 years of academic activity!
ATTRACTING GIRLS TO PHYSICS
Numerous Canadian academic institutions and non-profit
organizations are making efforts to generate interest in science and physics at an early age, preferably before secondary school. Such programs run year round or are struc-
SUMMARY
In recent years the overall climate for women
in academia in Canada has improved. Efforts
are being made to attract girls to science at a
young age. The enrollment of women across
undergraduate and graduate programs in
physical sciences has increased gradually in
the past decade, with a sharp increase at the
graduate level. In light of a large number of
upcoming retirements in academic positions, the presence of women in academia
will continue to grow, supported by efforts to
ensure equity in academia made by government agencies, academic institutions and
faculty associations.
tured as girls-only summer camps hosted by universities.
Activities are carefully selected to ensure that the participants have a large variety of opportunities to help them see
the connections between science and every day life, to
help the participants to gain confidence in their science
achievement, and ultimately to encourage their enrollment
in future science courses. The Canadian Association for
Girls in Science is an example of organization with chapters across the country that fosters early scientific literacy
through a variety of diverse, fun activities such as “the
physics of music, or the chemistry of cooking”. The
Techsploration program running in Nova Scotia is one
such example where role models well matched for the age
group of the students interact with the female students and
try to stimulate their interest in and enjoyment of science.
Academic units across Canada also support the local
schools in their efforts to attract girls to physics through a
variety of outreach programs such as science fairs. For
example, in Alberta girls represent slightly over 60% of
the number of participants in local science fairs. In the
Science Olympics part of the Regional Science Fair in
Alberta girls were up to 75 % of the numbers of participants! Clearly girls are interested in science and it is up to
us to design activities to generate and maintain their interest in physics and in science in general.
CLIMBING UP THE LADDER AND
NARROWING THE GENDER GAP IN
CANADIAN ACADEMIA
The Equity Review released in 2008 by the CAUT has
shown a firm upward trend in the enrollment of women in
higher education at the college, undergraduate and graduate levels. At the graduate level in physical sciences we
have experienced increases of 90% in the enrollment of
women compared with 1992. Regardless of the “feminization of universities”, the number of female students in
physical sciences still lags behind the number of male students. Aside from small regional differences, women continue to be under-represented in applied sciences.
Over the past 40 years the number of Canadian universities has increased from 30 to over 75. In parallel with this
steady growth, there has been an improvement in the status of women faculty across the country. Unfortunately,
A. Predoi-Crossa
<adriana.predoicross@
uleth.ca>, R. Austinb,
S. Bhadrac,
J. McKennad, L-H. Xue
and M. Steinitzf
aDept. of Physics and
of Lethbridge,
Lethbridge, AB
bPhysics Dept., Saint
Mary's University,
Halifax, NS
cDept. of Physics and
Astronomy, York
University, Toronto, ON
dDept. of Physics and
of British Columbia,
Vancouver, BC
eDept. of Physics,
University of New
Brunswick, Saint John,
NB
fDept. of Physics,
St. Francis Xavier
University, Antigonish,
NS
ICWIP 2008, KOREA
only 10 % of all women faculty members teach in applied
sciences, compared with a third of all men. While a sizable
difference in the proportions by gender at the same academic level remains [1], the difference has fallen dramatically and the trend appears to be continuing. However, at
the rank of full professor, and at the highest levels of university administration, women continue to be under-represented.
THE “MATERNAL WALL” IN CANADIAN
ACADEMIA
In spite of numerous efforts to reduce the gender differences in institutions across Canada, in recent years several feminist studies have pointed out that Canadian women
that have higher education may not encounter gender discrimination until they encounter the so called “maternal
wall” that hinders advancement in their professional
careers [2]. This is mostly because traditionally in our
country women are the ones who do most of the domestic
work, are in charge of household management, childcare
and elder care. The cumulative effect of all these factors is
that professional mothers simply are unable to find the
overtime hours that are often both expected and required
for advancement and success in their profession.
Professional mothers find themselves “mommytracked” [3] both financially and on the professional
advancement scale, with respect to their male counterparts. Sadly, it has been shown [4] that the pay gap between
young or middle-aged mothers and women of the same
age who have no children is now larger than the wage gap
between men and women from the same age group. In
Canada, organizations such as the Association for
Research on Mothering founded in Toronto at York
University, are making efforts to find strategies to help
mothers cope with the “maternal wall” in academia.
CONCLUSION
In recent years Canada has seen an increase in the number
of women at all academic levels in applied physical sciences. Empathy and a good understanding of all emotional and intellectual challenges faced by women in these disciplines will make the equity initiatives a success. The
trends observed in recent years will continue if the academic institutions and their faculty associations work
together with our government agencies with the goal of
obtaining equity at all levels in academia.
ACKNOWLEDGMENTS
Partial financial support for the Canadian team members
was provided by the Canadian Association of Physicists.
A. Predoi-Cross is grateful for financial support from the
Dean of Arts and Sciences and the Faculty Association of
University of Lethbridge. R. Austin acknowledges support
from Saint Mary's University. L.H. Xu acknowledges partial financial support from the University of New
Brunswick in Saint John. M. Steinitz acknowledges support from St. Francis Xavier University and the National
Research Council of Canada Research Press.
REFERENCES
1.
2.
3.
4.
266 C PHYSICS
A.R.R. Margulis, The Road to Success: A Career Manual - How to Advance to the Top, Academic Press, 2006.
Andrea O’Reilly, Rocking the Cradle, Toronto: Demeter Press, 2006.
H.A. Cummins, Women’s Studies International Forum, 28 222– 231 (2005).
Ann Crittenden, The Price of Motherhood: Why the Most Important Job in the World is Still the Least Valued, Holt Paperbacks,
2002.
IN
ICWIP 2008, KOREAAA
POSTER PRESENTED BY THE CANADIAN DELEGATION TO THE 3RD IUPAP
INTERNATIONAL CONFERENCE ON WOMEN IN PHYSICS - ICWIP2008
ENSEIGNEMENT
INTERNATIONAL PHYSICS OLYMPIAD, 2008
BY
ANDRZEJ KOTLICKI AND NATALIA KRASNOPOLSKAIA
S
imilarly to the competitions in Korea, Indonesia
and Singapore, this year’s Olympiad in Vietnam
was quite clearly an event of primary importance
to the Vietnamese government and educational
authorities. The president of Vietnam, his Excelleancy
Mr. Nguyen Phu Trong, and the Deputy Prime Minister
Prof Nguyen Thien Nhan, participated in the opening ceremony and stressed in their opening addresses the paramount importance of science, technology and education
for the development of Vietnam.
A Nobel Prize Laureate Prof. Jerome Friedman participated in the Olympiad activities, gave a lecture to the participants, and socialized with them.
In the experimental problem students had to measure the
efficiency of a solar cell. So 3 out of 4 problems (the third
theoretical problem was about Cherenkov radiation) had
something to do with “a green life style”.
Marking by the academic committee was very thorough
and fair and, in most cases, agreed closely with the marking of the leaders. The marking moderations (the process
of establishing the final mark acceptable by both leaders
and the local marking team) were performed in a good collegial atmosphere with very few real controversies.
Canada was represented by the following students:
Bo Cheng Cui (Bob) from BC
Jingyuan Zhang (Lynda) from Alberta
Jixuan Wang from Ontario
Keith Kaichung Ng from Ontario
Junjiajia Long (Bill) from Ontario
The social program was very entertaining and interesting,
with visits to monuments, temples and historical sites, an
excursion to the spectacular Halong Bay, and the continuous “flow” of excellent Vietnamese food.
The academic part of the
competition was organized
by the faculty members from
the
Hanoi
National
University of Education and
Institute of Physics and
Electronics,
Vietnamese
Academy of Science and
Technology.
Dr Andrzej Kotlicki
<[email protected].
ca>, Department of
Physics and
of British Columbia,
The team leaders were: Dr
Andrzej Kotlicki from the
Department of Physics and
Astronomy, University of
British Columbia, and Dr
Natalia Krasnopolskaia from
the Department of Physics,
University of Toronto.
The following 82 countries
were present at the 39th
International Olympiad:
The problems were very
interesting and well prepared. One of the theoretical
problems involved a modeling of the ancient waterpowered rice-pounding mortar. The other one involved
modeling the air flow in the
atmosphere and air pollution.
and
Dr Natalia
Krasnopolskaia,
Department of
Physics, University of
Toronto
Albania, Argentina, Armenia,
Australia, Austria, Azerbaijan,
Belarus, Belgium, Bosnia &
The opening ceremonies, with the Canadian team in the fore- Herzegovina, Brazil, Brunei,
Bulgaria, Cambodia, Canada,
ground.
Chile*, China, Colombia,
Croatia, Cuba, Cyprus, Czech
Republic, Denmark, Estonia, Finland, France, Georgia,
SUMMARY
Germany, Great Britain, Greece, Hong Kong, Hungary,
Iceland, India, Indonesia, Iran, Ireland, Israel, Italy, Japan,
The 38th International Physics Olympiad
Kazakhstan, Kuwait, Kyrgyzstan, Latvia, Liechtenstein,
(IPhO) was held in Hanoi, Vietnam from 20th
Lithuania, Macau, Macedonia, Malaysia, Mexico, Moldova,
to 19th of July, 2008. A total of 82 countries
Mongolia, Nepal, The Netherlands, Nigeria, Norway,
participated in the competition this year.
Pakistan, The Philippines, Poland, Portugal, Puerto Rico**,
Chile participated for the first time and Syria
Romania, Russia, Saudi Arabia, Serbia, Singapore, Slovakia,
sent an observer, planning to participate next
year. The team from Puerto Rico participated
unofficially as it does not represent an inde* new countries invited by the Organizing Committee to the
Olympiad this year
pendent country.
** invited by the Organizing Committee as a guest team.
268 C PHYSICS
IN
PHYICS OLYMPIAD 2008AA
The Canadian team receives their medals at the 2008 International Physics Olympiad held in Vietnam.
Slovenia, South Korea, Spain, Sri Lanka, Suriname, Sweden,
Switzerland, Syria*, Taiwan (Chinese Taipei), Tajikistan,
Thailand, Turkey, Turkmenistan, Ukraine, USA, Vietnam.
The best score (44.6 points) was achieved by Longzhi Tan
from China (Absolute winner of the 39th IPhO). The following limits for awarding the medals and the honorable
mentions were established according to the Statutes: Gold
Medal - 33 points (out of 50), Silver Medal - 26 points,
Bronze Medal - 21 points, and Honourable Mention 14 points. According to the limits, 46 Gold Medals,
47 Silver Medals, 78 Bronze Medals and 87 Honorable
Mentions were awarded. The list of the scores of the winners and the students awarded with honorable mentions
were distributed among all the delegations.
In addition to the regular prizes, the following special
prizes were awarded:
- for the best score (Absolute winner):
Longzhi Tan (China);
- for the best score in the theoretical part of the competition:
Longzhi Tan (China);
- for the best score in the experimental part of the
competition;
Yi-Shu Wei (Taiwan);
- for the best score among female participants:
Andrada Ianus (Romania):
- Gorzkowski Prize (for the best participant among
the countries that joined IPhO first in 2008):
Efraín Alfonso Pérez Argandoña (Chile);
- for the best Vietnamese competitor:
Huynh Minh Toan
The Canadian team performed very well, winning one
gold medal (Junjiajia Long) who was 6th in the world, two
silver medals (Jingyuan Zhang and Bo Cheng Cui) and
two bronze medals (Keith Kaichung Ng and Jixuan
Wang). It was the first time in the history of Canadian participation in the IPhO that all the Canadian team members
were awarded medals.
At the meeting of the International Board, the presidential
election was carried out according to a secret ballot.
Dr. Hans Jordens (The Netherlands) was elected the new
president.
At the end of the Olympiad, acting on behalf of the organizers of the next International Physics Olympiad, Dr. José
Luis Morán López, announced that the 40th International
Physics Olympiad will be held in Mérida, Mexico from
July 11th – 19th, 2009, and cordially invited all the participating countries to attend the competition.
CAP OFFICE
Mark your calendars:
June 7 – 10, 2009
CAP Congress, Moncton
CAP OFFICE - 2009 CONGRESS
The 2009 CAP annual Congress will be hosted by the University of
Moncton, located near some of the most scenic countryside in Atlantic
Canada. This will be an opportunity to celebrate the accomplishments of
Canadian Physicists from coast to coast and to enjoy the sights of beautiful
New Brunswick.
270 C PHYSICS
The Congress will begin on Sunday with plenary talks, division meetings, a
welcoming barbecue and a poster session. The Herzberg Memorial Public
Lecture will be held Monday night at the historic Capitol Theatre in
Moncton, and in addition to the CAP’s medal winners, plenary speakers
include Dava Sobel, author of “Galileo's Daughter”, a new biography of
Galileo based on recently recovered information from one of his daughters,
and Greg Flato of Environment Canada who will talk about the physics
involved in modeling climate change.
The Congress will continue through Wednesday afternoon with a variety of
invited and contributed sessions and special events, including a visit from
Isabelle Blain of NSERC, an exciting program is planned for the High
School Teacher’s Workshop, and the CAP Best Student Paper competition.
A lobster dinner will be served at the Congress banquet on Tuesday
evening.
We look forward to seeing you there!
For updates and program information, bookmark the Congress web
site at:
www.cap.ca/congress/congress.html
Abstract submission deadline:
IN
March 2, 2009
BUREAU DE L’ACP
Le Congrès annuel de l’ACP 2009 aura lieu à l’Université de Moncton, située
à proximité de certains des paysages les plus pittoresques du Canada atlantique. À cette occasion, nous célébrerons les réalisations des physiciens canadiens d’un océan à l’autre et profiterons des lieux majestueux du NouveauBrunswick.
Le congrès débutera le dimanche par des conférences plénières, des réunions
de division, un barbecue d’accueil et une session d’affiches. Lundi soir, la
conférence publique commémorative Herzberg sera présentée à l’historique
Théâtre Capitol de Moncton. En plus des récipiendaires des médailles de
l’ACP, nous accueillerons, entre autres conférenciers, Dava Sobel, auteure de
La Fille de Galilée (traduction de Galileo’s Daughter), une nouvelle biographie sur Galilée d’après la découverte récente de la correspondance avec l’une
de ses filles, et Greg Flato d’Environnement Canada qui s’entretiendra de la
physique impliquée dans la modélisation du changement climatique.
Le congrès se poursuivra jusqu’à mercredi après-midi avec une variété de sessions invitées et contribuées, ainsi que des événements spéciaux, dont la visite
d’Isabelle Blain du CRSNG, un programme captivant pour l’Atelier des
enseignants du secondaire et la compétition de l’ACP pour les meilleures communications étudiantes. Un repas au homard sera servi au banquet de l’ACP
mardi soir.
Au plaisir de vous voir!
Pour des mises à jour et des renseignements sur le programme, marquez
d’un signet le site Web du Congrès :
www.cap.ca/congress/congress-f.html
Date limite de soumission des résumés : 2 mars 2009
BUREAU DE L’ACP - CONGRÈS 2009
À vos calendriers:
7 au 10 juin 2009
Congrès de l’ACP, Moncton
CAP OFFICE
CAP OFFICE - 2009 CONGRESS
BEST STUDENT PRESENTATIONS /
MEILLEURES COMMUNICATIONS ÉTUDIANT(E)S
272 C PHYSICS
The CAP is hosting The Best Student Oral Presentation
at Congress and The Best Student Poster at Congress
competitions at the 2009 CAP Congress. There will be
cash prizes awarded by the CAP for the top three overall oral presentations by a student and for the top three
overall poster presentations by a student. In addition,
various divisions and sponsors will hold more focused
competitions. The intent of all these competitions is to
encourage graduate students to present their research
work to, and interact with, the Canadian physics community at an early stage in their careers. Students
should present the work and take primary responsibility for the content of the presentation as well as the written abstract and, if selected as a winner, the extended
abstract for Physics in Canada.
Dans le cadre de son congrès de 2009, l'ACP tient les
concours Meilleure communication orale étudiante au
congrès et Meilleure affiche étudiante au congrès. Elle
décernera des prix en argent pour les trois meilleures
communications orales globales et pour les trois
meilleures affiches globales par un étudiant. De plus, une
variété de divisions et commanditaires tiendront des concours plus ciblés. L'idée de tous ces concours est d'encourager les étudiants diplômés à présenter leurs travaux
de recherche à la collectivité canadienne de la physique
et à interagir avec elle en début de carrière. Les étudiants
doivent présenter leur communication et assumer la
responsabilité première de son contenu, ainsi que le
résumé soumis et, si choisi comme gagnant(e), le
résumé élargi pour La Physique au Canada.
Undergraduate students who fulfill these criteria are
also welcome to present.
Les étudiants de 1er cycle qui répondent à ces critères
sont aussi les bienvenus.
In order to be considered for this competition, a student
must be a member of the CAP. Membership is free for
undergraduates as well as for the first year as graduate
member. Students can become a member or renew
their membership online at http://www.cap.ca/mem/
mem.html
Pour être candidat à ce concours, l'étudiant doit être
membre de l'ACP. L'adhésion est gratuite pour les étudiants de 1er cycle et durant une année pour les étudiants
diplômés. Vous pouvez adhérer ou renouveler votre
adhésion en ligne à http://www.cap.ca/mem/mem-f.html
At the time of submission of their abstract, students
must specifically indicate their desire to participate in
the competition by selecting the appropriate option.
The abstract preference will determine whether you are
registering for the oral paper or poster competition. A
fuller description of both competitions on the CAP
Congress website.
A student entering either the overall best paper, best
poster, or a non-division sponsored competition will
automatically be entered into the appropriate divisional
competition if one exists and if he/she is a member of
that division. The following divisions are offering prizes:
Atomic and Molecular Physics and Photon Interactions
Condensed Matter and Materials physics
Instrumentation and Measurement Physics
Medical and Biological Physics
Optics and Photonics
Particle Physics
Plasma Physics
Theoretical Physics
SPONSORED COMPETITIONS
In addition to the competitions noted
above, Atomic Energy of Canada
Limited is sponsoring prizes for the
best student presentation relating to nuclear engineering, or reactor and radiation physics (both oral and
poster). Details can be found on the CAP’s Congress
website.
IN
Au moment de soumettre son résumé, l'étudiant doit indiquer spécifiquement son choix de participer au concours
de communications orales ou d’affiches de l'ACP en choisissant l’option appropriée. La préférence indiquée dans
le résumé déterminera s’il veut s’inscrire au concours oral
ou par affiches. Une description plus complète des deux
concours se trouve sur le site internet du congrès de
l’ACP.
Un étudiant qui se présente au concours meilleure communication orale, meilleure affiche, ou à une compétition
commanditée par un autre organisme que les divisions
de l’ACP sera automatiquement inscrit au concours de la
division appropriée, si celle-ci en a un et si l’étudiant est
membre de la division. Les divisions qui suivent remettront des prix:
Physique atomique et moléculaire et d'interaction avec
les photons
Physique de la matière condensée et des matériaux
Physique des instruments et mesures
Physique médicale et biologique
Optique et photonique
Physique des particules
Physique des plasmas
Physique théorique
COMPÉTITIONS COMMANDITÉES
En plus des concours indiqués cidessus, Énergie atomique du Canada
Limitée remettra des prix pour les
meilleures communications étudiantes en génie nucléaire ou en physique des rayonnements et des réacteurs (orale et affiche). Les détails se
trouvent sur le site internet du congrès de l’ACP.
LIVRES
BOOK REVIEW POLICY
Books may be requested from the Book Review Editor, Richard Hodgson, by using the online book request form at http://www.cap.ca.
CAP members are given the first opportunity to request books. Requests from non-members will only be considered one month after the distribution date of
the issue of Physics in Canada in which the book was published as being available (e.g. a book listed in the January/February issue of Physics in Canada will
be made available to non-members at the end of March).
The Book Review Editor reserves the right to limit the number of books provided to reviewers each year. He also reserves the right to modify any submitted
review for style and clarity. When rewording is required, the Book Review Editor will endeavour to preserve the intended meaning and, in so doing, may find
it necessary to consult the reviewer. Beginning with this issue of PiC, the text of the book reviews will no longer be printed in each issue, but will be available
on the CAP website.
LA POLITIQUE POUR LA CRITIQUE DE LIVRES
Si vous voulez faire l’évaluation critique d’un ouvrage, veuillez entrer en contact avec le responsable de la critique de livres, Richard Hodgson, en utilisant le
formulaire de demande électronique à http://www.cap.ca.
Les membres de l'ACP auront priorité pour les demandes de livres. Les demandes des non-membres ne seront examinées qu'un mois après la date de distribution du numéro de la Physique au Canada dans lequel le livre aura été déclaré disponible (p. ex., un livre figurant dans le numéro de janvier-février de la
Physique au Canada sera mis à la disposition des non-membres à la fin de mars).
Le Directeur de la critique de livres se réserve le droit de limiter le nombre de livres confiés chaque année aux examinateurs. Il se réserve, en outre, le droit de
modifier toute critique présentée afin d'en améliorer le style et la clarté. S'il lui faut reformuler une critique, il s'efforcera de conserver le sens voulu par l'auteur
de la critique et, à cette fin, il pourra juger nécessaire de le consulter. Commençant par cette revue de PaC, le texte des critiques de livre ne
sera plus imprimé dans chaque revue, mais sera disponible sur le page Web de l’ACP.
BOOKS RECEIVED / LIVRES REÇUS
The following books have been received for review. Readers are
invited to write reviews, in English or French, of books of interest to
them. Books may be requested from the book review editor,
Richard Hodgson by using the online request form at
http://www.cap.ca.
Les livres suivants nous sont parvenus aux fins de critique. Celle-ci
peut être faite en anglais ou en français. Si vous êtes intéressé(e)s à
nous communiquer une revue critique sur un ouvrage en particulier, veuillez vous mettre en rapport avec le responsable de la critique
des livres, Richard Hodgson par internet à http://www.cap.ca.
A list of ALL books available for review, books out for review, and
copies of book reviews published since 2000 are available on-line -see the PiC Online section of the CAP's website :
http://www.cap.ca.
Il est possible de trouver électroniquement une liste de livres
disponibles pour la revue critique, une liste de livres en voie de
révision, ainsi que des exemplaires de critiques de livres publiés
depuis l'an 2000, en consultant la rubrique "PiC Électronique" de la
page Web de l'ACP : www.cap.ca.
GENERAL INTEREST
ELECTRICAL TRANSPORT IN NANOSCALE SYSTEMS, Massimiliano Di
Ventra, Cambridge University Press, 2008; pp. 476; ISBN: 978-0-52189634-4 (hc); Price: $80.00.
MODERN QUANTUM FIELD THEORY: A CONCISE INTRODUCTION,
Thomas Banks, Cambridge University Press, 2008; pp. 267; ISBN:
978-0-521-85082-7 (hc); Price: $65.00.
ON SPACE AND TIME, A. Connes, M. Heller, S. Majid, R. Penrose, J.
Polkinghorne, A. Taylor, Cambridge University Press, 2008; pp. 287;
ISBN: 978-0-521-88926-1 (hc); Price: $26.00.
UNDERGRADUATE TEXTS
BOSE-EINSTEIN CONDENSATION IN DILUTE GASES, CJ. Pethick and
H. Smith, Cambridge University Press, 2008; pp. 569; ISBN: 978-0521-84651-6; Price: $80.00.
FUNDAMENTALS OF PLASMA PHYSICS, Paul M. Bellan, Cambridge
University Press, 2008; pp. 609; ISBN: 978-0-521-52800-9 (pbk);
Price: $75.00.
INTRODUCTION TO QUANTUM THEORY, Harry Paul, Cambridge
University Press, 2008; pp. 176; ISBN: 978-0-521-87693-3 (hc);
Price: $50.00.
GRADUATE TEXTS
AND PROCEEDINGS
NONEQUILIBRIUM QUANTUM FIELD THEORY, Estaban A. Calzetta and
Bei-Lok Hu, Cambridge University Press, 2008; pp. 533; ISBN: 9780-521-64168-5 (hc); Price: $90.00.
PLASMA PHYSICS AND FUSION ENERGY, Jeffrey P. Freidberg,
Cambridge University Press, 2008; pp. 667; ISBN: 978-0-521-73317-5
(pbk); Price: $80.00.
BOOK REVIEWS
BOOK REVIEWS / CRITIQUES DE LIVRES
Book reviews for the following books have been received and posted to the Physics in Canada section of the CAP’s website :
http://www.cap.ca. Review summaries submitted by the reviewer are included; otherwise, the full review can be seen at the url listed with
the book details.
Des revues critiques ont été reçues pour les livres suivants et ont été affichées dans la section “La Physique au Canada” de la page web de
l’ACP : http://www.cap.ca. Les résumés des critiques de livre sont inclus tels que soumis; toutefois, la critique complète peut être lue au
lien url indiqué avec les détails du livre.
CANADA’S FIFTY YEARS
IN SPACE,
Gordon Shepherd and Agnes Kruchio, Apogee
Books, 2008, pp. 280, ISBN 978-1-894-959-728; Price: CAN$26.95 [To read detailed review,
please see http://www.cap.ca/brms/reviews/
Rev933_638.pdf]
This book is a fascinating story of Canada’s first
half century in space, the interesting people
involved, and what made them click. It covers the
personal stories of space pioneers Balfour Currie,
Frank Davies and Don Rose and their research
work on the upper atmosphere; the support from
the US Air Force (USAF) in ionospheric physics,
geomagnetism, cosmic radiation and auroral
physics research, and how this support allowed
space research to take root in Canada; the phenomenal growth of space science in the 60’s due
in part to the visions and foresights of individuals; the interval of transition in the 70’s and the
80’s, and the effects of the “Chapman report” and
the Space Shuttle program in this period; and
developments in the two decades since the creation of the Canadian Space Agency, including
the Canadian Astronauts Program and Canadian
scientific instruments on several international
satellite missions. The book is full of interesting
stories, anecdotes, first-person accounts and reminiscences. Its perceptive analyses on what made
people of scientific visions and entrepreneurial
spirits click – and what works and what doesn’t
in science research – serve as a good source of
inspiration for science researchers, policy makers, and students alike.
Andrew Yau
University of Calgary
Calgary, Alberta, Canada
CLASSICAL MECHANICS, R. Douglas
Gregory, Cambridge University Press, 2006, pp.
596, ISBN 0-521-82678-0 (hc); 0-521-53409-7
(pbk); Price $120.00/60.00.
This textbook would serve as an excellent companion to a student learning any level of undergraduate classical mechanics and a thorough
knowledge of 1st year calculus. Topics are introduced at a fundamental level, avoiding all the
glossy pictures and handwaving arguments presented in most 1st year textbooks. Each topic is
presented in a thorough, clear, and self contained
manner which never leaves the reader feeling
lost. I also feel that the author has spent just the
right amount of time on each topic, providing
sufficient examples and problems without getting
bogged down in lengthy discussions. This textbook will be specifically useful for students
274 C PHYSICS
IN
learning the subject for the first time because
solutions are given to every problem in the book.
Parts 1 and 2 as well as Chapters 16 and 17 of
Classical Mechanics constitute a second year
classical mechanics course. Chapter 1 is an introduction to vector algebra and contains many
interesting problems in geometry that can be
solved using vector techniques. Chapters 2-4 and
16-17 give a standard introduction to kinematics
and dynamics, including rotating reference
frames. An excellent coverage of linear oscillations and normal modes, including damped
motion, forced motion, coupled oscillations are
given in chapter 5. This is followed up by the
general theory of small oscillations in chapter 15.
Chapter 6 and part 2 form good coverage of energy/momentum principles of particles and rigid
bodies. There are also two well written chapters
on orbits in a central field (chapter 7) and nonlinear oscillations (chapter 8).
Part 3 of this book combined with chapter 19 and
certain sections of the chapters described above
are suitable for a 3rd year advanced mechanics
course. Topics covered include conservation
principles, the calculation of variations,
Lagrange’s equation, Hamilton’s principle as
well as their applications to rigid body dynamics.
Steven Conboy
CLIMATE CHANGE - BIOLOGICAL AND
HUMAN ASPECTS, Jonathan Cowie.
Cambridge University Press, New York, 2007,
pp. xvi + 487; ISBN: 978- 0-521-69619-7
(ppbk), 978- 0-521-87399-4 (hdbk); US$52.00/
120.00. [To read the detailed review, please see
http://www.cap.ca/brms/reviews/Rev879_607.
pdf]
The author of this remarkable book is an
extremely articulate and erudite writer as readers
sense as soon as they have read a few pages. The
perspective is unusual as the author is a biologist
by training rather than a physical scientist and
this has been broadened by his experience in representing the Institute of Biology to both the UK
government, the public and school children in
particular. This innovative book starts with a
brief introduction to climate change and this is
followed by a careful discussion of the various
indicators of past climates. Two chapters cover
the periods before and after the EoceneOligocene extinction some 35 mya. This is followed by an examination of the present climate
and biological change and an informed discus-
sion of the warming now occurring and its likely
future impact. The role of population growth,
energy supply, human health, food security in
causing the present warming are then discussed
followed by the role that biology can play in
reducing this warming. The role various UN
organized conferences have played in getting to
the Kyoto Accord is explained. A careful analysis
of energy sustainability and future energy policy
options leads to a discussion of future human and
biological change and the various alternative
solutions. This outstanding book should be on the
must read list of every scientist.
Harvey A. Buckmaster,
Adjunct Professor of Physics
University of Victoria, Victoria, BC
EINSTEIN: HIS LIFE AND UNIVERSE,
Isaacson, W. , New York: Simon& Schuster.,
2007, pp. 675, ISBN 978-0-743-3264730 (hc),
$39 [To read the detailed review, please see
http://www.cap.ca/brms/reviews/Rev906_626.
pdf]
W. Isaacson’s Biography of Albert Einstein, the
most famous scientist of the 20th century, should
be a book assigned to every curious student of
science, either in high school, college or university. The style of the book, the accuracy and
lucidity of the science descriptions, as well as the
thorough attention to historical detail and to the
roots of Einstein’s ideas, make the book capture
the reader’s imagination and makes it hard to put
down. Using a wealth of historical material, scientific papers, including some of the newly available documents from the recently opened
archives, Isaacson creates a portrait of Einstein
that captures his scientific genius, as well as his
human personality, with all its triumphs, contradictory political views, flaws, and personal struggles. Scientifically presented, but understandable
to a first year science (or even a high school) student, the development of Einstein’s theories of
special and general relativity, and the theory of
the photoelectric effect, is placed in historical
context, thus allowing the reader to trace
Einstein’s thought process, as well as to see the
roots of Einstein’s difficulties with Quantum
Mechanics and its probabilistic nature.
Independent of the science background of the
reader, the book will be able at the same time to
satisfy and ignite readers’ curiosity and make
them feel a part of one of the most amazing scientific revolutions of the 20th century, and the
person who started it all. Isaacson’s biography of
Einstein will also be of great interest to curious
CRITIQUES DE LIVRES
lay people outside of academia and to anybody
who ever wondered about the universe we live in.
Dr. Marina Milner-Bolotin
Department of Physics, Kerr Hall East 329E
Ryerson University
ELECTRON CORRELATION IN METALS,
Kosaku Yamada, Cambridge University Press,
2004; pp. 245; ISBN: 0-521-57232-0 (hc); Price:
$100.00.
This book is an attempt to describe the physics of
strongly correlated electron systems, which has
played an important role in several phenomena in
condensed matter physics, most notably in magnetic and superconductive properties of metals.
The aim of the author was to describe the theory
of electron correlations based on the Fermi liquid
theory. There are nine chapters in the book covering a brief introduction of the Fermi gas
(Chapt.1) with standard treatment of the freeelectron system, and exchange, screening, etc.,
Fermi liquid theory (Chapt. 2), screening effect
of an impurity charge in metals, (Chapt. 3),
Anderson’s orthogonality theorem, and the
Kondo effect (Chapt. 4). The other chapters deal
with magnetic impurities in metals (Chapt. 5),
and a lot of technical details about the Hubbard
Hamiltonian (Chapt. 6). The chapter on heavy
fermions (Chapt. 7) deals with the formal details
that might be helpful for practitioners. Chapter 8
deals with the transport theory. Here Hall conductivity, optical conductivity, etc. are treated by
the Fermi liquid theory. The final chapter provides a description of high-temperature superconductivity in terms of the Fermi liquid theory.
My overall assessment is that the book is highly
technical with brief and dry texts interspersed
among the equations. The treatment of the topic
is largely theoretical, with a few instances where
connections to experiments were made. The book
might interest the experts as useful reference
material, but because of the complete absence of
detailed introductory texts in each chapter, nonexperts will find it rather difficult to follow.
Tapash Chakraborty
University of Manitoba
INTRODUCTION TO THE ELECTRON
THEORY OF METALS, U. Mizutani,
Cambridge University Press, 2000; pp. 576;
ISBN: 0-521-65248-0; Price: US$80. [Review
by Tapash Chakraborty, Univ of Manitoba; To
read the detailed review, please see http://
www.cap.ca/brms/reviews/Rev733_560.pdf]
LINEAR ELASTIC WAVES, John G. Harris,
Cambridge University Press, 2001, pp. 158,
ISBN: 0-521-64383 (ppk), U.S. $25. [To read
the detailed review, please see http://www.cap.
ca/brms/reviews/Rev293_623.pdf]
The book Linear Elastic Waves by J.G. Harris is
intended as an introductory text for graduate students in the physical sciences. A vast knowledge
of elasticity is not required, but the reader should
be familiar with a related field theory such as
electromagnetic theory. Moreover, the student
must have an extensive background in calculus,
differential equations, and complex analysis.
The material covered includes an outline of the
general model equations needed to understand
linear elastic waves. The physics and mathematics behind wave reflection and refraction, surface
and guided waves, as well as edge diffraction and
dispersion are also included. Additionally, it was
shown that the kinematical form of any wave can
be constructed from a collection of plane waves.
As well, a discussion of the integral representations of solutions to rather general problems in
elastic-wave propagation is presented.
Throughout the book, care was taken to develop
the required mathematics, with proofs provided
in many places. A full appreciation of this material, however, requires that the reader to have an
understanding of mathematics at the level of a
senior undergraduate physics student. In particular, complex analysis is used extensively in the
last two chapters on elasticwave radiation and
guided waves, respectively.
To help the reader gain a further understanding of
the covered material, the author has provided references to other relevant texts throughout the
book. In addition, approximately 40 practice
problems have been included which require the
reader to extend the concepts beyond the specific cases and conditions that are presented within
this text. In general, these problems are relatively difficult and helpful hints have been provided
for just a few of them.
Overall, Linear Elastic Waves is a great book and
is suitable as a reference text for postgraduate
physics students as well as practicing
researchers.
Lance Parsons
Memorial University, St. John’s
Newfoundland, Canada
I would have a difficult time recommending this
book in its present edition for the reason that it
has not been thoroughly proofread. This shows in
both grammatical errors (at least one per page)
and factual errors that are likely typos.
Grammatical errors are annoying (particularly in
these numbers), but would not in themselves prevent a textbook from being useful. Examples of
factual errors include reference to incorrect isotopes (eg, C-40 rather than C-14), text that is
inconsistent with the equations or figures it
describes, and digits out of place (eg, stating the
regulatory annual nuclear worker dose limit
1000x higher than actual). Before using this
book as a textbook, or reference text, I would like
to see another edition that has been scrubbed for
oversights – it wouldn’t be difficult and would
greatly improve the usefulness of what is otherwise a thorough overview of radiation detection
fundamentals.
Eva Marczak
Toronto, Ontario
THE IDEAS OF PARTICLE PHYSICS: AN
INTRODUCTION FOR SCIENTISTS, Guy D.
Coughlan, James E. Dodd, Ben M. Gripalos,
Cambridge University Press, 2006, pp. 254,
ISBN 0-521-67775-0 (pbk); 978-0-743-3264730
(hc), $50/$100. [To read the detailed review,
please see http://www.cap.ca/brms/reviews/
Rev844_618.pdf]
The book The Ideas of Particle Physics by
Coughlan, Dodd and Gripaios promises to be an
introduction aimed to scientists who do not necessarily have a physics background. The book is
primarily focused on the general ideas of particle
physics, the recent findings and the basic physics
behind the phenomenon.
The author intended for this book to provide a
good basis for understanding the fundamental
principles of radiation detection and their effective use in detection technologies. This book
could be a text for a university level course on
techniques of radiation detection and a reference
for incorporating radiation detection effectively
in experimental apparatus.
The book starts from very basic ideas in physics
such as the concept of matter and light, special
relativity, fundamental forces, quantum mechanics etc. In later chapters, the basics of particle
physics are addressed: strong and weak interactions, gauge theory, deep inelastic scattering,
quantum chromodynamics, theory of quarks and
electron positron collisions. At the end it all
comes together with a discussion of the standard
model. The format of the book is different from
standard text books. There are not any derivations and proofs. The number of equations is
minimal and there is hardly any assumed specialized knowledge required. There are many illustrative graphs throughout the book and there are
no end-of-chapter problems.
The first two chapters review the types and
sources of radiation and its interaction with matter. Gas, liquid, solid state, scintillation and photo
detectors are then reviewed. Use of detection systems is extensively addressed in chapters on signal processing, statistics for data analysis, data
analysis software and data acquisition systems. A
chapter on dosimetry, biological effects of radiation, and radiation protection basics is included.
Mathematical concepts are illustrated through
sample problems with worked solutions throughout the book, highlighted in shaded boxes to distinguish them from the text. There are an additional 5-20 practice problems (without solutions)
at the end of each chapter. An extensive bibliography is also provided at the end of each chapter.
I liked that the chapters were short and not too
detailed. It was an easy read and ideas were well
explained without referring to many equations
and calculations. An interested reader should
consult the text books for further details. The
book is divided into ten “parts”, each containing
several chapters. Each chapter covers a topic in
less than ten pages and starts with one or two
introductory paragraphs. I found it useful that
although the ten parts and chapters are connected
to each other, each is still self-contained. What I
would have liked to see is a chapter or two dedicated to the experiments, how they are designed,
performed and what are the challenges. There are
many parts in the text where the authors refer to
various experiments and recent results; however
PHYSICS AND ENGINEERING OF RADIATION DETECTION, S. Ahmed, Academic Press
(Elsevier), 2007, pp: 800, ISBN-13:978-0-12045581-2, ISBN-10:0-12-045581-1; Price:
$95.00 US. [To read the detailed review, please
see http://www.cap.ca/brms/reviews/Rev892_
617.pdf]
BOOK REVIEWS
I think it would have been more complete to
mention the way experiments work with the same
introductory touch as the rest of the book.
The authors claim is well delivered and the book
suits a scientist reader. Obviously more knowledge in physics would be useful and would make
the book a much easier book to read. However as
long as the reader is familiar with how science
works the book would be useful. In my opinion
this book is well suited for physics graduate students in fields other than particle physics who
want to learn the general ideas and not the
details. I am a graduate student with a research
field out side of particle physics. I recommend
this book to undergraduate students in physics,
graduate students with no particle physics background and other scientists who are seeking a
general knowledge of this exciting field of
physics. This book is definitely not a popular science book and the reader is required to have a
scientific understanding of physics. After reading
this book it will be much easier to follow more
thorough text books.
Sanaz Vafaei
THE SPECTRA AND DYNAMICS OF
DIATOMIC MOLECULES: REVISED AND
ENLARGED EDITION, H. Lefebvre-Brion and
R. W. Field, Elsevier, 2004. pp: xxx+766, ISBN
0124414567(sc); Price $111. [To read the
detailed review, please see http://www.cap.ca/
brms/reviews/Rev832_546.pdf]
The spectra and dynamics of diatomic molecules
is a considerable expansion of Perturbations in
the spectra of diatomic molecules [i]. As the
authors note, the use of the word “perturbations”
in the title of the earlier work resulted in a rather
limited use and appreciation of its content. Many
interesting dynamical processes in molecules are
related to perturbations in their spectra. The
expansion that formed the present work has clarified this.
Where I find this work disappointing is in the
book that deserves to be widely read.
organization and incorporation of the new mateIain R. McNab,
rial into the re-worked chapters – particularly in
Lash Miller Chemical Laboratories,
Chapters Three and Six. In the earlier work
University of Toronto
spherical tensors were avoided in considering
problems of angular
momentum; in the presUNIVERSITY OF TORONTO
ent work spherical tensor algebra is someDEPARTMENT OF PHYSICS
times
used.
Faculty Position in Biological Physics
Unfortunately, the new
material, which uses
The Department of Physics at the University of Toronto is pleased to announce
spherical tensors, has
the search for a tenure stream appointment in theoretical, experimental or computational Biological Physics at the rank of Assistant Professor, with a starting date
been added into the
of July 1, 2009 or shortly thereafter.
older text haphazardly
and the result is a mess.
We seek candidates with a Ph.D. in Physics or a related field, and with proven or
For readers not already
potential excellence in both research and teaching. We are particularly interested
familiar with spherical
in the general area of complex systems, including genomics, proteomics, neurotensor algebra, Chapters
science, systems biology and applications of statistical mechanics and nonlinear
dynamics to biological systems, although outstanding candidates in any field of
Three and Six will be
biological physics are encouraged to apply. The new appointment will have the
largely opaque. The
opportunity to interact with existing groups in biological physics and related areas
Spectra and Dynamics
of nonlinear physics, quantum optics and condensed matter physics. In addition,
of Diatomic Molecules
the University of Toronto is home to one of the largest and most active biomedis
a
considerably
ical research communities in North America. We invite prospective candidates to
expanded version of its
visit our home page at www.physics.utoronto.ca. The salary will be commensupredecessor. The two
rate with qualifications and experience.
additional chapters are
Applications, including a curriculum vitae and a summary of proposed research
well written and informshould be sent to:
ative, and there is much
Professor Michael Luke, Chair
here that can only be
Department of Physics, University of Toronto
found elsewhere in
60 St. George Street
research papers (if at
Toronto, ON, Canada M5S 1A7
all). I wish that the
email: [email protected]
authors had chosen to
take
a
consistent
Three letters of reference should also be sent directly to the above address under
separate cover. Applications will be reviewed beginning December 1, 2008 until
approach to angular
the position is filled. Those received by December 1, 2008 will be given first conmomentum theory in
sideration.
the re-worked chapters,
rather than the mixed
The University of Toronto offers the opportunity to teach, conduct research, and live, in one of
the most diverse cities in the world. The University of Toronto is strongly committed to diversiavoidance and use of
ty within its community, and especially welcomes applications from visible minority group
spherical tensor algebra
members, women, Aboriginal persons, persons with disabilities, members of sexual minority
that is now present.
groups and others who may contribute to further diversification of ideas. All qualified candiNevertheless, this is an
dates are encouraged to apply; however, Canadians and permanent residents will be given
priority.
excellent and useful
This “revised and enlarged” work contains two
completely new chapters (One and Nine) while
Chapters Two through Eight are revised and
expanded versions of Chapters One through
Seven of Perturbations in the spectra of diatomic
molecules. The earlier book was reviewed by
R.N. Dixon [ii]; his comments are still pertinent
and will not be repeated here. The two new chapters (Chapter One – Simple spectra and standard
experimental techniques, Chapter Nine Dynamics) are excellent additions. Chapter One
is simple introductory material. Chapter Nine
expounds the link between dynamical concepts
and spectral perturbations. In chapter nine we
move into the time domain with descriptions of
the evolution of diatomic molecule states, and of
time-domain experiments with which such
motions can be probed. Many of the new and
most exciting techniques in spectroscopy are
mentioned in this chapter, such as photoassociation spectroscopy and the use of crafted laser
pulses for control of molecular dynamics. Here
too we are introduced to some of the theoretical
techniques with which dynamics can be treated in
quantum mechanics. The chapter concludes with
a discussion of the dynamics of polyatomic molecules.
276 C PHYSICS
IN
AFFICHAGES DE POSTES
Tenure-Track Faculty Position
Theoretical Condensed Matter Physics
Department of Physics
McGill University
The Department of Physics at McGill University invites applications for a tenuretrack position at the rank of Assistant Professor, beginning as early as September
2009. The appointment will be in the area of Theoretical Condensed Matter Physics.
The applicant will be expected to become a member of the Centre for the Physics of
Materials, which includes faculty members from the departments of Physics and
Chemistry as well as research scientists in industrial laboratories. The focus of the
Centre is on research at the boundary between Condensed Matter Physics and
Materials Science. Faculty members are currently active in nanoscience, nonequilibrium materials, biophysics, quantum information theory, surface science, magnetism,
and strongly-correlated electronic systems. The Centre has extensive computer facilities that include a state-of-the-art Beowulf cluster, and benefits from access to
McGills CLUMEQ Supercomputer Centre. One of its major strengths is the extensive interaction and collaboration that exists between theory and experiment.
The department has active groups in Astrophysics, Biophysics, Condensed Matter,
Nuclear, Particle, and Theoretical Atmospheric Physics. For more information about
McGill and the Department of Physics, consult our home page.
We welcome applications from candidates in all major areas of condensed matter
physics, with a proven record of excellence in research and also the capacity for
excellence in teaching. Applicants should submit a detailed curriculum vitae and a
statement of teaching interests as well as a research plan. They should also arrange
for three letters of reference to be sent directly to:
Professor Charles Gale, Chair
Department of Physics, McGill University
3600 rue University
Montréal, QC Canada, H3A 2T8
Review of applications will begin December 29th, 2008. The successful candidate
will be supported by a generous start-up package and could be nominated for a
Canada Research Chair.
All qualified candidates are encouraged to apply; however Canadian citizens and permanent
residents of Canada will be given priority. McGill University is committed to equity in employment.
c
i
Ryerson University is known for innovative programs built on the integration of theoretical and practically oriented learning. More than 95 undergraduate
and graduate programs are distinguished by a professionally focused curriculum and strong emphasis on excellence in teaching, research and creative
activities. Ryerson is also a leader in adult learning, with the largest university-based continuing education school in Canada.
TENURE-TRACK FACULTY POSITION – Department of Physics
The Department of Physics offers undergraduate and graduate programs, and is currently composed of fifteen full-time faculty and
six staff members. The Department has a core group of scientists who have secured substantial external peer-reviewed funding
for cutting-edge research in Medical Physics and Biomedical Engineering. Our faculty members collaborate extensively with the
surrounding biomedical community in what the City of Toronto has designated as the Discovery District, home to seven worldrenowned hospitals and more than 30 specialized medical and related sciences centres. The Department is also engaged in research
in the field of Physics Education. More information on the Department of Physics can be found at www.ryerson.ca/physics.
Faculty of
Engineering,
Architecture and
Science
The Department invites applications from outstanding candidates for a tenure-track faculty position at either the Associate or
Assistant Professor level. The focus of this search is on Biomedical Applications of Radiation. Applicants must have a strong background in physics, possess
an earned doctorate in Physics or a related field, and demonstrate excellence in research and teaching. The position is subject to budgetary approval, and
can begin as early as July 2009.
Interested candidates should submit a current CV and statements of proposed research directions and teaching interests, and should arrange for at least three
letters of reference to be sent directly, to: Dr. Pedro Goldman, Chair, Department of Physics, Ryerson University, 350 Victoria Street, Toronto, Ontario, Canada,
M5B 2K3. This posting will remain open until the position is filled, but the Department Appointments Committee will start reviewing applications on January 5, 2009.
Ryerson University has an employment equity program and encourages applications from all qualified individuals, including Aboriginal peoples, persons with disabilities,
members of visible minorities and women. Members of designated groups are encouraged to self identify. All qualified candidates are encouraged to apply; however,
Canadians and permanent residents will be given priority.
E N G I N E E R I N G
I
A R C H I T E C T U R E
I
S C I E N C E
ALL UNDELIVERABLE
COPIES IN
CANADA /
TOUTE
CORRESPONDANCE NE
POUVANT
ETRE
LIVREE AU
CANADA
should be
returned to /
devra être
retournée à
:
Canadian
Association of
Physicists/
l’Association
canadienne des
physiciens et
physiciennes
Suite/bur. 112
Imm. McDonald
Bldg.
Canadian Publications Product Sales Agreement No.
40036324 / Numéro de convention pour les envois de
publications canadiennes : 40036324
Univ. of/
d’Ottawa,
150 Louis
Pasteur, Ottawa,
Ontario
K1N 6N5

ASTROPHYSICS / ASTROPHYSIQUE

Transcription

Documents pareils

Chapitre 4 vitesse de la lumière

Notice biographique Repères biographiques communs Nom

Le concours d`orthographe du Balfroid

M2R ASEP SYLLABUS des COURS UE1: Rayonnement et gravitation

Plan du camping - Camping Les Flots Bleus

La Provence du 05/04/09 - Observatoire de Haute

A la manière de Claude Roy…

TP Informatique

Vendredi 11 décembre Décorations d`étoiles pour le sapin