Precision radial velocity surveys for exoplanets [Elektronische Ressource] / put forward by Mathias Zechmeister

English
120 Pages
Read an excerpt
Gain access to the library to view online
Learn more

Description

DissertationsubmittedtotheCombinedFacultiesoftheNaturalSciencesandMathematicsoftheRuperto-Carola-UniversityofHeidelberg,GermanyforthedegreeofDoctorofNaturalSciencePutforwardbyDiplom-PhysikerMathiasZechmeisterbornin:Wippra,GermanyndOralExamination:2 February2011PrecisionRadialVelocitySurveysforExoplanetsReferees: Prof.Dr.ThomasHenningProf.Dr.JoachimWambsganß1AbstractSincethe90sastronomershavediscoveredaround500extrasolarplanets.Mostofthemhavebeenfoundwith the radial velocity method. In this work we present our precision radial velocity measurementsfor a sample of 40 M dwarfs and 30 solar-like stars. The data sets originate from four different instru-ments (UVES, CES+LC, CES+VLC, and HARPS) and are investigated for indications of planets. Weperform several statistical tests for excess-variability, long-term trends, and periodicities. For the latterpurpose, we have developed further a commonly used period analysis tool, the so called Lomb-Scargleperiodogram.Our radial velocity precision of a few m/s is approximately sufficient for the aspired goals, namelythe search for terrestrial planets in the habitable zones of M dwarfs and the search for Jupiter analoguesaroundsolar-likestars.Wedemonstratethiswithmassupperlimits.Ourdataanalysisdoesnotrevealanynew planet, while we can confirm the two known planets around the solar-like stars ι Hor and HR 506as well as the long-term trend forε Ind A.

Subjects

Informations

Published by
Published 01 January 2011
Reads 8
Language English
Document size 17 MB
Report a problem

Dissertation
submittedtothe
CombinedFacultiesoftheNaturalSciencesandMathematics
oftheRuperto-Carola-UniversityofHeidelberg,Germany
forthedegreeof
DoctorofNaturalScience
Putforwardby
Diplom-PhysikerMathiasZechmeister
bornin:Wippra,Germany
ndOralExamination:2 February2011PrecisionRadialVelocitySurveys
forExoplanets
Referees: Prof.Dr.ThomasHenning
Prof.Dr.JoachimWambsganß1
Abstract
Sincethe90sastronomershavediscoveredaround500extrasolarplanets.Mostofthemhavebeenfound
with the radial velocity method. In this work we present our precision radial velocity measurements
for a sample of 40 M dwarfs and 30 solar-like stars. The data sets originate from four different instru-
ments (UVES, CES+LC, CES+VLC, and HARPS) and are investigated for indications of planets. We
perform several statistical tests for excess-variability, long-term trends, and periodicities. For the latter
purpose, we have developed further a commonly used period analysis tool, the so called Lomb-Scargle
periodogram.
Our radial velocity precision of a few m/s is approximately sufficient for the aspired goals, namely
the search for terrestrial planets in the habitable zones of M dwarfs and the search for Jupiter analogues
aroundsolar-likestars.Wedemonstratethiswithmassupperlimits.Ourdataanalysisdoesnotrevealany
new planet, while we can confirm the two known planets around the solar-like stars ι Hor and HR 506
as well as the long-term trend forε Ind A. Moreover, we were able to identify several binaries and one
browndwarf.OurresultsareinagreementwithestimatesforthefrequencyofJupiter-likeplanetswhich
isaround1%forMdwarfsand10%forsolar-likestars.
Zusammenfassung
Seitden90-zigerJahrenhabenAstronomenetwa500extrasolarePlanetenentdeckt.DiemeistenvonIh-
nenwurdenmitderRadialgeschwindigkeits-Methodegefunden.IndieserArbeitpräsentierenwirunsere
präzisen Radialgeschwindigkeitsmessungen für ein Sample von 40 M-Sternen und 30 sonnenähnlichen
Sternen. Die Datensätze stammen von vier verschieden Instrumenten (UVES, CES+LC, CES+VLC,
und HARPS) und werden auf Hinweise von Planeten untersucht. Dazu führen wir statistische Tests
zu Exzessvariabilitäten, Langzeittrends und Periodizitäten durch. Für letzteren Zweck, haben wir ein
häufiggenutztes Werkzeug für die Periodenanalyse, das sogenannte Lomb-Scargle Periodogramm, wei-
terentwickelt.
Unsere Radialgeschwindigkeitsgenaukeit von wenigen m/s ist annähernd ausreichend für die ange-
strebten Ziele, nämlich die Suche nach terrestrischen Planeten in der habitablen Zone von M-Sternen
und die Suche nach Jupiter-artigen Planeten um sonnenähnliche Sterne. Wir demonstrieren dies mit
oberenMassengrenzen.AusunsererDatenanalysegehtkeinneuerPlanethervor,währendwirdiebereits
bekannten Planeten für die sonnenähnlichen Sterne ι Hor und HR 506 sowie den Langzeittrend für
ε Ind A bestätigen können. Darüber hinaus konnten wir einige Doppelsterne und einen Braunen Zwerg
identfizieren. Unsere Ergebnisse stimmen mit den Erwartungen für die Häufigkeit von Jupiter-artigen
Planetenüberein,welcheetwa1%fürM-Sterneund10%fürsonnenähnlicheSterneist.2Contents
Abstract 1
1 Introduction 5
1.1 Methodsforexoplanetdetection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6
1.1.1 Directimaging . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6
1.1.2 Thetransitmethod . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6
1.1.3 Astrometry . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
1.1.4 Gravitationalmicrolensing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
1.1.5 Othermethods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
1.2 Theradialvelocitymethod . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10
1.2.1 Measuringpreciseradialvelocities . . . . . . . . . . . . . . . . . . . . . . . . . 11
1.2.2 Selfcalibrationwithagasabsorptioncell . . . . . . . . . . . . . . . . . . . . . 12
1.2.3 Simultaneouscalibrationwithemissionlamps . . . . . . . . . . . . . . . . . . . 13
1.2.4 Othertechniques . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14
1.3 Outlineofthiswork . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14
2 ThegeneralisedLomb-Scargleperiodogram 15
2.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15
2.2 ThegeneralisedLomb-Scargleperiodogram(GLS) . . . . . . . . . . . . . . . . . . . . 16
2.3 NormalisationandFalse-Alarmprobability(FAP) . . . . . . . . . . . . . . . . . . . . . 18
2.4 EquivalencesbetweentheGLSandSigSpec . . . . . . . . . . . . . . . . . . . . . . . . 20
2.5 ApplicationoftheGLStotheKeplerianperiodogram . . . . . . . . . . . . . . . . . . . 21
2.6 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26
2.7 Accompanyingauxiliarycalculations . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27
2.7.1 DerivationofthegeneralisedLomb-Scargleperiodogram(GLS) . . . . . . . . . 27
2.7.2 VerificationofEq.( 2.19) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28
2.8 CommentsontheGLSperiodogram . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30
2.8.1 ErrorestimationfortheGLSparameters . . . . . . . . . . . . . . . . . . . . . . 30
2.8.2 Apolargrid . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33
2.8.3 TheF-distribution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34
3 TheMdwarfsurveywithESOVLT+UVES 37
3.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37
3.2 TargetsandObservations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39
3.3 Dataanalysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39
3.3.1 Secularacceleration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39
3.3.2 Testsforvariabilityandtrends . . . . . . . . . . . . . . . . . . . . . . . . . . . 43
3.3.3 Periodogramanalysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47
3.3.4 Upperdetectionlimits . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47
3.3.5 CorrelationbetweenRVandHα index? . . . . . . . . . . . . . . . . . . . . . . 50
3.4 Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 54
34 CONTENTS
3.5 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 54
3.6 Accompanyingauxiliarycalculations . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55
3.6.1 Relationbetweenindexandequivalentwidth . . . . . . . . . . . . . . . . . . . 55
3.6.2 ResponseoftheGLSperiodogramwhenaddingasinewave . . . . . . . . . . . 55
3.7 AdditionalnoteregardingtheplanetdiscoveryforGJ433 . . . . . . . . . . . . . . . . . 58
4 TheESOCESandHARPSsurvey 61
4.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 61
4.2 Thesample . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 62
4.3 Instrumentsanddatareduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 64
4.3.1 CES+LongCamera . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 64
4.3.2 CES+VeryLongCamera . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 64
4.3.3 HARPS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 66
4.3.4 DetailsoftheRVcomputationfortheCES+VLCdata . . . . . . . . . . . . . . 67
4.3.5 CombiningtheLCandVLCdata . . . . . . . . . . . . . . . . . . . . . . . . . 68
4.3.6theCESandHARPSdata . . . . . . . . . . . . . . . . . . . . . . . 69
4.4 Analysisoftheradialvelocities . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 69
4.4.1 Excessvariability . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 71
4.4.2 Long-termtrends . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 74
4.4.3 SearchforPeriodicitiesandKeplerianorbits . . . . . . . . . . . . . . . . . . . 75
4.4.4 Orbitalsolutions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 78
4.4.5 Testsontheresidualsofthecompanionhostingstars . . . . . . . . . . . . . . . 80
4.4.6 Detectionlimits . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 81
4.5 SummaryandConclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 82
4.6 Plotsofallradialvelocitytimeseries . . . . . . . . . . . . . . . . . . . . . . . . . . . . 84
4.7 AccompanyingTables . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 90
4.8 Plotsofallperiodograms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 94
4.9 Plotsofalldetectionlimits . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 99
5 SummarisingConclusions 103
5.1 Outlook . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 104
5.1.1 PrecisionRVmeasurementsinthenear-infrared . . . . . . . . . . . . . . . . . . 104
5.1.2 Newcalibrationsources . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 105
5.1.3 Stellarnoiseasthelimitingfactor? . . . . . . . . . . . . . . . . . . . . . . . . . 106
Bibliography 107
Acknowledgement 115Chapter1
Introduction
Discovering planets and new worlds has long been a dream of humankind inspiring people to undertake
new enterprises. Overcoming old fashioned, anthropocentric world views like the geocentric system or
seeing the Sun or the Milky Way as centres of the universe, the idea was at hand that planets may exist
aroundotherstars.
Thehistoryuntilthefirstdiscoveryofanexo-planetisquiteeventfulandalsoaccompaniedbymany
falsedetections,e.g.therewereearlyclaimsofplanetdetectionsaround61Cygby Strand(1943,1957)
or around Barnard’s star by van de Kamp (1963) via astrometry. Later on however, they turned out to
be spurious detections (Heintz, 1978). Struve (1952) proposed to use the radial velocity (RV) method
to search for planets and outlined that it would be possible to find close-in, massive Jupiters with the
precisionavailableatthattime(∼200m/s).
Campbell et al. (1988) noted for the binary star γ Cep (K0IV): “probable third body variation of
25m/s amplitude, 2.7yr period”. Their precision RV measurements indicated probably for the first time
a bona fide exoplanet. Later this team argued that stellar phenomena (rotational modulation of active
regions) might be a more plausible explanation than a planetary companion (Walker et al., 1992), but
finally Hatzes et al. (2003) reinforced the planetary hypothesis (see also Walker 2008 for a review).
Noteworthy is also the RV detection of the companion to HD 114762 (F9V) with a planetary minimum
mass (11M ) by Latham et al. (1989). However, the authors themselves pointed out that this object isJup
probablyabrowndwarfwhichwaslatercorroboratedbyanestimateofthestellarrotationaxis(Cochran
etal.,1991).
The first, widely accepted planet detection was around the pulsar PSR B1257+12 by Wolszczan &
Frail (1992). It was an amazing discovery for four reasons: (i) The planet host is a neutron star. (ii) It is
1a planetary system with two planets . (iii) The planet masses are only a few Earth masses. (iv) It was a
veryluckydetection.“Thanks”toamechanicaltrackingproblemofthe305mAreciboradiotelescopein
PuertoRico,Wolszczanandhisgroupwereallocatedtimeforaprogrammetosearchforpulsarsoffthe
galacticplane.Indeed,theyfoundnewpulsarsandcollectedanunusuallylargenumberofpulsetime-of-
arrival observations to characterise them. The data set for PSR B1257+12 could not be explained with a
pulsarmodelalone,butwhenaccountingfortwoorbitingplanetsthesignalwaswelldescribed.Sothese
planets were found as a by-product of a programme that did not aim to search for planets (Wolszczan,
private communication, manuscript for Astronomy and Astrophysics Review in prep., editor Lissauer).
So far only a second pulsar (PSR B1620-26, Thorsett et al. 1993; Backer et al. 1993) is known to host a
planet.Theprogressinthisfieldofpulsartimingishamperedbytherequirementsofextensivedatasets
andlargeradiotelescopesinordertofindplanetswiththismethod.
Nowadays, the detection of 51 Peg b by Mayor & Queloz (1995) is mostly referred to as the first
exoplanet discovery. It is the first planet found around a solar-like star (G5V) and was immediately
verified by an American group ( Marcy & Butler, 1995; Marcy et al., 1997). 51 Peg b has an orbital
periodofjust4.2dandaminimummassof0.47M .Thiswasevidencethatclose-inJovians,i.e.thoseJup
planets proposed by Struve (1952) for search programmes, indeed exist. From then on, many similar
1Laterathirdplanet(Wolszczan,1994)andfourthunconfirmedobjectwereannouncedinthissystem.
56 CHAPTER1. INTRODUCTION
2MASSWJ1207334−393254
778 mas N
55 AU at 70 pc
E
Figure 1.1: Examples for direct imaging detections. (Left) Planet to a brown dwarf (Chauvin et al.,
2004). (Right) Putative planets of the solar-like star GJ 758 with a scale for the solar system (Thalmann
etal.,2009).
objects have been found and belong to a class called hot Jupiters. These systems are very different from
our solar system where Jupiter has an orbital period of 12.3yr. The method, which Mayor & Queloz
(1995) employed, was the radial velocity technique that measures the Doppler shift introduced by the
gravitationalpullofanorbitingcompanion(Sect.1.2).
2In October 2010, the exoplanet list counts∼500 planets and planet candidates which were discov-
ered by the different methods briefly explained in Sect. 1.1. So far, the most successful technique is the
RV method which has revealed more than 350 exoplanets. The RV technique is the basic method in this
thesisandisexplainedinmoredetailinSect.1.2.
1.1 Methodsforexoplanetdetection
1.1.1 Directimaging
It is very challenging to directly image a planet orbiting around a star. One has to overcome the huge
brightnesscontrastandthesmallangularseparationbetweenthebrightstarandthefaintplanet.Several
techniques, such as coronography, nulling interferometry, or angular differential imaging, have been
developedtosolvethetechnicaldifficulties.
To enhance the detection probability astronomers have focussed on young, nearby systems, because
in their early stages planets are relatively bright compared to their host stars. Indeed, some groups have
successfully imaged some wide giant planets. Chauvin et al. (2004) imaged for the first time directly a
5M -massplanetaccompanyingabrowndwarf(Fig.1.1).AnotherexampleisthesubstellarcompanionJup
of the K dwarf GQ Lup which might have a planetary mass (Neuhäuser et al. , 2005). Further exciting
detections are the planet around the A3V star Fomalhaut (Kalas et al., 2008) and the planetary system
to the A5V star HR 8799 (Marois et al., 2008). GJ 758 b is the first planet-like object imaged to a sun-
likestar(Thalmannetal.,2009)inaseparationofonly29AUandwithatemperatureofonly600K.For
HR8799cJansonetal.(2010)demonstratedthatevendirectspectroscopyispossible.However,themass
estimates rely often on age estimates based on evolutionary models and, e.g., new asteroseismologic
measurements indicate an older age for HR 8799 and imply that the imaged companions are brown
dwarfs(Moyaetal.,2010).
1.1.2 Thetransitmethod
The principle of the transit technique is well known, because it has been applied since almost hundred
years to study eclipsing binary systems, and with today’s photometric precision it is also applicable to
2
http://exoplanet.eu/