EVOLUTION of CLASSICAL RADIO
SOURCES
PART I 1
1 Types of radio sour es 1
1.1 FRII typeradio sour es . . . 3
1.2 SizeofFRIItyperadiosour esfromGigahertz-Peaked-Spe trumSour es to GiantRadio Galaxies . . . 3
1.3 Cy les of the a tivity . . . 4
2 Theoreti al basis 5 2.1 Physi alpro esses inthe FRII typesour es . . . 5
2.2 Energy losses . . . 6
2.3 Sour e's sizes vs. other physi alfa tors . . . 8
3 Analyti al models of FRII type radio sour es 9 3.1 Sour e dynami s . . . 9
3.2 Sour e energeti s and luminosity evolution . . . 10
3.3 Determiningdynami alageand otherphysi alparametersforreal FRII type radio sour es . . . 12
3.4 Limitationsof the KDA model . . . 13
3.4.1 Asymmetry of radio sour es . . . 13
3.4.2 Unlimitedexternal density prole . . . 13
3.4.3 Limitationfor the jet a tivity . . . 14
PART II 15 4 Asymmetries of FR II type radio sour es predi ted with the KDA model 15 4.1 Lobe's length and total luminosity asymmetries . . . 15
4.2 Asymmetries resultingfrom dierent
β
exponent . . . 154.3 Asymmetries resultingfrom dierent axialratio
R
t
. . . 184.4 Asymmetries resultingfrom dierent
α
inj
parameter . . . 204.5 Asymmetries resultingfrom dierent
k
′
parameter . . . 224.6 Dis ussion and on lusions . . . 24
4.6.1 Modelpredi tions for the asymmetries . . . 24
4.6.2 Modelpredi tion in omparisonwith observationaldata. . . 25
5 The observational onstraint for the model of the radio-jets propaga-tion through the X-ray haloIGM interfa e (Kuligowska et al. 2009) 26 5.1 The base of the revised G-KW model . . . 27
5.3.1 Sample1 (GRGsample) . . . 31
5.3.2 Sample2 (Distant-sour e sample) . . . 31
5.3.3 Sample3 (3CRR sample) . . . 32
5.4 Ageing analysis of the samples'sour es . . . 36
5.5 Results of the modeling. . . 37
5.6 Observational onstrain of the model . . . 37
5.6.1 ComparisonoftheModel'sPredi tionwiththeObservationalData 38 5.6.2 Ageand physi al parameters of the samplesour es . . . 42
5.7 Dis ussion of the resultsand on lusion . . . 42
6 Modi ationof KDAmodelto itsversion appli able for radiosour es with non- ontinuous a tivity 51 6.1 The argument for anextension of the basi KDA model . . . 51
6.2 Extension of the original KDA model . . . 51
6.2.1 Generalbasis of the extended model . . . 51
6.2.2 Adiabati evolution of the o oon inthe ase of terminated nu- lear a tivity . . . 52
6.2.3 Spe tralageing inthe syn hrotrontheory. . . 53
6.2.4 The analyti alformulaforintegration of radiopower . . . 54
6.2.5 Predi tions of the extended model. . . 55
6.2.6 Dis ussion . . . 60
PART III 65 7 Appli ation of the extended model to a few sele ted radio sour es 65 7.1 Sample of examined radiosour es . . . 65
7.2 Fitting pro edure . . . 69
7.3 Appli ation of dierent dynami almodels to the sample sour es . . . . 70
7.4 Dis ussion and nal on lusions . . . 77
Even after a few de ades of resear h and investigations, the dynami al evolution of
powerfulradiogalaxiesisstillfarawayoffullunderstanding.The lassi aldynami alapproa h
toades riptionoftheintera tionbetweenthe"light"relativisti jets, arryinganoutstanding
amount ofenergy fromtheA tive Gala ti Nu leus(AGN)andintera ting withtheexternal
gaseous environment, annotprovide an uniquesolutionfor a numberof detailedparameters
des ribing the underlied physi al pro esses.The problem is that besides the observed linear
sizeofasour e(althoughproje tedonthesky)that anbeinterpretedasalengthofthejets,
theabove pro essesresult intime-dependent shape of theradiospe trumof agiven sour e.
All existinganalyti al models of dynami al evolution of lassi al radio sour es arebased
onsome un ertain assumptionsabout,for example: 1) initial power-law energy spe trum,2)
equipartition ofenergybetween therelativisti ele trons andmagneti eld,3) possible
rea - elerationoftheseele trons,4)parti ipationofprotonsintheenergylosses.However,re ently
thesemodels anbepartly onnedbyX-raymapping,unfortunatelyonlyfornearbyandthe
brightest sour es. In spite of the above, the existing models suer other severe de ien ies.
All of these models assume: 1) exponential prole of the ambient environment (simplied
King's model, King 1972), whi h is valid for a limited external halo and is not justied for
thelargest(thusold)radiogalaxies, 2) ontinuousinje tionofenergy viathejetsof onstant
powerthatisevidentlynot the asefortheobserved,so alleddouble-doubleradiostru tures
inwhi htwopairsoftwinlobesareobservedandinterpretedbyaterminationoftheprimary
jeta tivityand the appearan e ofa new, se ondary a tivity.
In this Thesis I addresssome of the above issues and modify thebasi analyti al model
of Kaiser, Dennet-Thorpe and Alexander, 1997 (hereafter referred to as KDA), as well as
ompareobservedspe traofafewexemplaryradiosour eswithpredi tionsofboththeKDA
modeland itsmodiedversions.
Theoutlineof theThesis isasfollows:Se tion 1summarize dierentapparent
morpholo-gies of radio sour es and dene the type of sour es for whi h analyti al models of their
dynami alevolution is onsidered here. Se tion 2 outlines physi al pro esses resulting inthe
observed morphology, brightness distribution, and spe tral shape of the sour es. Se tion 3
des ribesthe existing dynami almodels ofradio galaxy'sevolution and itslimitation due to
the real examples of observed radio galaxies. Se tion 4 presents the appli ation of original
KDAmodelin aseofstudyingthe observedasymmetriesofFRIItyperadiosour es.Se tion
5 is dedi ated to the revision of the existing model, assuming that radio sour es evolve in
two-media environment. Se tion 6 des ribes an attempt to extend the lassi al KDA model
ofthesour e'sevolutionwithfor assumed ontinousnu lear a tivitytoitsgeneralization
ap-pli able to the ase of sour es withnon- ontinous inje tion of relativisti ele trons. Finally,
Se tion7presentsappli ation oftheextendedmodelto afewrealradiogalaxieswithsteeply
1 Types of radio sour es
All existing radio sour es an be dened as elestial obje ts that emit rather strong
radio waves. However, they are basi allydivided into two general ases. The rst ase
in ludesallof "normal"galaxies inwhi h the dominantradiationinthe radiodomain
results from the star-formation pro esses, supernovae remnants (e.g. Cassiopeia A),
and the existen e of louds of ionized hydrogen (HII regions). The angular extent of
the sour e in this ase does not ex eed a solid angle of the opti al galaxy itself. The
se ond ase onsistsof the galaxieswith the dominantradio emissionoriginatingfrom
the nu lear a tivity - so- alled lassi al radio sour es. Their opti al spe tra are not
the simple superpositions of the spe tra of parti ular stars in the galaxy, but show
also hara teristi emissionlinesdue tothe high-energypro esses takingpla e intheir
enters. Su ha lassi alradiosour ehas linear,elongated,so- alled"double" stru ture
extendingusually far outside the parent galaxy.
Classi al radio sour es had been divided by Fanaro and Riley (1974) into two
morphologi altypes: FRI and FRII. This lassi ationin based onthe observed
mor-phology of the large-s ale radio stru tures. The FRI type radio sour es are brighter
towards its enter, whileFRIIsour es are thebrightest atthe edgesortheir two radio
lobes. The two FR types are also broadly separated by the luminosity; FRII sour es
aremoreluminousinthe radiowavelengthsthan FRItypes. Theoriginaldividingline
waspla edat
5 × 10
25
W Hz
−
1
atobservingfrequen y of178 MHz.Thisdivision isdue
tothe e ien y of the energy transportinto the sour e's radio stru ture - the entral
engines of FRII sour es supply the lobes with the energeti parti les more e iently
than inthe FRI ase. The jets of the latter appear tobe subsoni lose tothe nu leus
and radiate asigni ant amount of their total energy on their way to the radio lobes.
In turn, the jets in FRII sour es remain relativisti on their entire length, and the
so- alledhot spotsvisibleontheendsofthe radiolobes an beinterpretedasthe
man-ifestation of supersoni sho ks formed when the jet abruptly terminates en ountering
the external intergala ti mediumof some density and pressure.
Figures(1) and (2)show ar hetypesof the abovetwo morphologies:radio galaxies
3C405 (Cygnus A) of FR II type and 3C31 of FRI type. The ontent of this thesis is
ourtesyof NRAO/AUI.
Figure 2: Example of FR I type radio sour e: 3C31 radio map at 5 GHz (VLA). Image
FR II type radio sour es are omposed of double, ellipsoidal,radio-emitting regions
(so- alled radio lobes) situated more or less symmetri ally on the both sides of their
host galaxieswhi hinmost ases are largeellipti algalaxies ontainingsupermassive
bla k holes (SMBH) of masses
10
6
− 10
9
M
⊙
intheir enter. Theselobe stru turesareoriginated from twin jets outowing in opposite dire tions from the A tive Gala ti
Nu leus(hereafterreferred toasAGN).The radioemissionof these lobesresultsfrom
theele tronseje ted atnearlythe speedoflightthroughalongjetfromthe oreofthe
galaxy and deposited in radio lobes. The ele trons are trapped by the magneti eld
around the galaxy and produ e radio waves. In most of the powerful observed FR II
sour estheir jets end insigni antly brighter regions situatedatthe ends of the lobes
-hot spots learly seen in Figure(1).
The most famous example of FR II type radio sour es is Cygnus A. It has been
dis overedbyGroteReberin1939.CygnusAisoneof thestrongest radiosour esseen
from the Earth and, along with Cassiopeia A and Puppis A, had been the rst radio
sour e identied with opti al sour e (Baade and Minkowski, 1954) and, so, the rst
even observed lassi al radiogalaxy.
1.2 Size of FR II type radio sour es from
Gigahertz-Peaked-Spe trum Sour es to Giant Radio Galaxies
FRII type radio galaxies show a large variety of their linear sizes. The size of radio
galaxiesisdenedhereasthelineardistan ebetweenthebrightedgesofthebothradio
lobes. These sizes range from less then 100 p - Gigahertz-Peaked-Spe trum (GPS)
sour es up to about 4 Mp - Giant Radio Galaxies (GRG). Between these extreme
ases we observe a lot of medium size radio galaxies. Their typi al size is about 100
to400 kp . In general, this size is dependent of the age of the sour e - the older it is,
the more time the jets had to a t forming a large radio stru ture. On the other side,
the sour e's large size may resultfrom the lo allyo urred lowdensity ofthe ambient
mediumin whi h the sour e evolves( f. Se tion 3.1).
GPS sour eshaveamaximumof uxdensityintheirradio spe traatabout1GHz
andlinearsizes notex eeding 1kp .Theyare onsideredtobeyoung,re ently formed
sour es. Several studies show that they probably expand during their time evolution
and transform to Compa t Steep Spe trum (CSS) sour es with linear sizes of about
10 - 15 kp . Then, in a next stage of the evolution, they transform into the phase of
typi alFR I of FR II type radio sour e. In a spe i ondition they an evolve tothe
ultimatephase,i.e.GRGs.Theyaredenedasradiosour eswithlinearsizesofatleast
1megaparse (foraxedHubbleConstant
H
0
=0.71km s
−
1
Mpc
−
1
andthe osmologi al
parameters
Ω
m
= 0.27
andΩ
Λ
= 0.73
(Λ
CDM model). GRGs are interesting be ause of the fa t that their existen e brings out important questions about their evolution,galaxies what suggests they an be the nal stage of sour e's evolution.
Until re ently the largest radio galaxywas 3C236that features a "double-double"
morphology onsisting of thegiantreli of
∼
4.4Mp size and aninnerstru ture of∼
2 kp size whi h is a young, ompa t steep spe trum sour e. (Willis, Strom & Wilson1974.) However, in 2005 the larger radio galaxy have been dis overed. J1420-0545 has
itslinearproje tedsizeof4.69Mp (Ma halskietal.2008).Despiteofthis,J1420-0545
isalso interesting due to itsrelatively high redshift (
z = 0.3067
).An analysis of the rea hed sour es' size (and their age) in the fun tion of their
temporary radio luminosity was onsidered by Kaiser & Best (2007). A summary of
their onsiderationsis given in Se tion2.2.
1.3 Cy les of the a tivity
The a tivity of radio galaxies results from the pro esses of a retion of a matter
onthe bla k hole with magneti eld and angularmomentum. This a tivity evidently
annot be innite; it an be stopped aftera time dened as "jet ut-o" time. It was
estimatedtobenolongerthenabout
10
8
years(Komissarov&Gubanov1994). When
the a tivity ends even temporarily, the lobes are not longer powered with the inow
of relativisti parti les, soas the result, the luminosity of the radio stru ture tends to
ease. This ee t is observed mostly on higher radio frequen ies (
ν
>1000 MHz) and thisee tseemstobe ommoninthe aseofold,large-sizedradiosour es.Inastudyoftheradiostru turesof somegiantradiogalaxies,Subrahmanyan, Saripalli&Hunstead
(1996) drew attention to a variety of morphologi al features in the GRGs whi h were
indi ative of interrupted or episodi nu lear a tivity. It implies that GRGs may had
attained their large sizes as a result of restarting of their nu lear a tivity in multiple
phases.
Observations also indi ate the existen e of some spe i ase of radio sour es
-so alled Double - Double Radio Galaxies (DDRGs). Their extended radio stru tures
onsist of the inner, younger double stru tures with almost pure power-law spe tra
and the outer, larger, steepened spe trum stru ture, onsidered to be the older one.
Observations suggest alsothat su h double-double stru tures predominantly o ur in
the ase of the largest radio galaxies. However, only a small fra tion of GRGs has
double-double stru tures (S hoenmakers et al. 2000). Studies of the possible nu lear
re urren e in radio sour es are in their early days and the role of su h a re urrent
2.1 Physi al pro esses in the FRII type sour es
The radio galaxy's a tivity onne ted with jets formation results fromthe pro esses
of a retionof a matteronthe super massive bla k hole(hereafter referredto SMBH)
withmagneti eld andangularmomentum.AGNsare hara terizedby ahigh rate of
the a retion of gaseous matter situated lose to the entral SMBH. A reted matter
formsso- alleda retiondis swheretherotationalenergyoftheSMBHistransformed
into plasma'sinternal energy, and then, radiatedinthe wide range of the
ele tromag-neti spe trum. Magneti eld of the SMBH links the a retion dis with the plasma
pla ed outside the ergosphere. That implies that the high fra tion of the plasma an
be a eleratedin the dire tiondened by the spin of the AGN.
Jets (and double radiostru tures) are formed onlyin some fra tionof allobserved
AGNs. Wilson & Colbert (1995) and Blandford (1999) suggest that jets of so- alled
radio loud AGNs are only produ ed in the ase of entral SMBH with very strong
magneti eld and high spin. High values of the spin results probably from the
merg-ing event of two SMBH (Merrit & Ekers, 2002.) Physi al pro esses leading to the
observed stru ture of FR II type radio sour e are determined by so- alled "Standard
AGN model"(Blanford, Rees &S heuer 1974).The s hemati diagramofthe sour e's
omponentsis shown on Figure(3).
the host galaxy, then the medium of the luster ontaining that galaxy, and nally
it may rea h the Intergala ti Medium (hereafter IGM) whose density an be higher
than the density of the jet. Intera tion of the jet with those media produ es a bow
sho k.Thissho k formaninterfa ebetween thesho ked andtheunperturbed external
medium.A ompressionof relativisti jetmaterialhaspla e athot spots,the brightest
regionsof the ends of the jets where external mediumintera ts with the jet,resulting
in medium's parti les a elerations after their transition through the bow sho k. The
jet'spowerdepends onthetotal AGNrotationalenergy.The higherthis energyis,the
fastertherelativisti parti lesofthejetsmove.IftheAGNisrelativelyweak,theradio
lobesare powered with poorly ollimatedjets in whi h turbulent for es o ur and the
velo ity of the relativisti parti lesslows down.
2.2 Energy losses
The FR II type sour es have predominantly power-law spe tra. Non-thermal
on-tinuum radio emission of the lobes is due to both syn hrotron pro ess and
inverse-Comptons atteringofambientphotons ofthe osmi mi rowaveba kground(CMBR).
Thesyn hrotronradiationarisesfromultrarelativisti hargedparti les(with Lorentz
fa tor of
γ ∼ 10
4
) thatintera twith the magneti eld.This emissionin highly
polar-izedand anbeobserved withintheentireele tromagneti spe trum,soit anbeeasily
re ognized.A ording to lassi alele trodynami sthe life-timeofrelativisti ele trons
emitting mediumand short radio waves has to be relatively short. Be ause the radio
lobes are stable, long-living stru tures (asthe observations indi ate), they have to be
ontinuously powered with the amountsof new parti les. The sour eof theseparti les
an beonly the entral partof the galaxy -AGN.Due toso- alledsyn hrotron losses,
themostenergeti parti lesloosetheirenergy inthe fastestrate.Itimpliesthat,inthe
absen e of the onstant inje tion of the new parti les, syn hrotron losses in the radio
lobes result a violent steepening of the observed radio spe tra of the sour es. Most
energeti ele trons preferentially emit at high frequen ies, so the high-frequen y part
of the spe trum rapidly steepens and nally uts o.
The Inverse Compton ee t originates from the intera tion of the jet's relativisti
parti les with CMBR photons. A ording to Kardashev (1962), this pro ess need to
be taken into a ount in the analysis of the sour e's total energeti s, as well as the
adiabati expansions ofthe lobes(resulting inboth energylosses and gains),Coulomb
losses (o urring when relativisti parti les ollide with thermal ele trons of a given
density), and energy a quisition by systemati or sto hasti a eleration. It had been
proved that Coulomb losses play a major role in total energy losses only at very low
energies(
γ < 10
).Bothradiativeandadiabati lossesare dominantathigherparti le's energies, espe iallyforγ > 10
3
(Murgia 1996).
L
D
L − D
reprodu ed inFigure (4).
Figure4:The FRII type sour e's luminosityevolution of alobe.
The sour e's luminosityis linked to astage of the itslife and the dominant energy
losses at a this stage. For the ase of dominant adiabati loses, the radiation losses
ould be negle ted. In this regime the radio luminosity depends on sour es size as
L ∝ D
(8−7β)/12
. It implies that at this stage the expe ted luminosity an de rease or in rease depending on whetherβ >
8
7
orβ <
8
7
, respe tively. The syn hrotron losses are the most important at the early time of sour e's life be ause the energy densityof the magneti eld in whi h parti les radiates de rease while the sour e grows and
grows older. The radio luminosity in this regime is the onstant fun tion of its size
L ∝ D
0
∼ const
. The external density prole is now approximated by a power law
and the relation between the length D and the sour e age t depends on the value of
thepower-lawexponentβ.Inthis regimesyn hrotronlosses dominateovertheInverse
Compton ee t.
As the magneti eld de reases due to the ageing of the sour e, Inverse Compton
lossesremain onstantand nallydominateoversyn hrotronlosses.Atthisphasevery
strong totalenergy losses are observed
L ∝ D
(−4−β)/(5−β)
Thelobe'slengthisnotonlythesimplefun tionofitsage.Thetotalsizedepends also
on amount of energy released from the AGN and the properties of the IGM inwhi h
thesour eevolves. ForexampleGRGsareonlyafra tionof 6
%
ofthetotalpopulation of luminous radio galaxies (Laing, Riley & Longair, 1983). It may imply that theirenormous sizes result from the "density voids" in the IGM. Additionally, only about
10
%
ofknownGRGslieatredshiftlargerthan0.5.Themaximumlinearsizeofobserved radio galaxies de reases asD ∝ (1 + z)
−
3
(Gopal-Krishna & Wiita 1987). This fa t
is in a goodagreement with the theory assuming that in the adiabati allyexpanding
Universe, lled up with hot, uniform IGM, its density in reases as
̺
IGM
∝ (1 + z)
3
,
(Kapahi1989) anditskineti temperaturegrows up proportionallyto
T
k
∝ (1 + z)
2
.It
impliesthat the pressure of IGM hanges as
p ∝ (1 + z)
5
(Cotter1998). The pressure
of the external medium likely de ides how far the radio stru ture an expand. That
explain why most of the GRGs are situated lose to us. However, nding GRGs at
mu hhigherredshiftmay provethatparts oftheIGM, withsigni antlylowerdensity
then an average one, are alsoin early osmologi al epo hs. Forthis reason GRGs are
usefultoolfor studyingtheevolutionofradiosour esitself,aswellasthe osmologi al
All analyti al models take into a ount known physi al pro esses o uring in FR II
type radio sour es. They hara terize these sour es in terms of their dynami s and
energeti s onne tedtothe luminosityevolution.Allof themoriginatefromthe
"stan-dard model" fordouble radio sour es (e.g.Blandford &Rees 1974)in whi h lobesare
omposed of a sho ked jet and IGM material.The axiallength of a lobe results from
thebalan ebetween the jet'sthrustand theram pressureof theexternalmedium,and
itswidth is afun tion of the jet's parti lesinternal pressure.
3.1 Sour e dynami s
This balan eisa base forthe sour e's dynami sdes ribed by the moresophisti ated
analyti al model of Kaiser & Alexander (1997; hereafter referred to KA). The jets
thrust,
Π
jet
,and itsbalan ewith theram pressure ofthe gaseousenvironmentisgiven byΠ
jet
∼
Q
jet
v
jet
≈ ρ
a
v
h
2
A
h
,
(1)where
Q
jet
andv
jet
arethejet'spowerandspeed;thelateroneisassumedtobe loseto thespeedoflight(v
jet
≈ c
).ρ
a
istheambientdensity,v
h
isvelo ityofthejet'shead( f. Figure(3)), andA
h
isthe ross-se tion areaof the bowsho k there the time-averagedΠ
jet
is dis harged over.Transforming Eq.(1) we have
v
h
=
d
dt
r
j
≈
Q
jet
ρ
a
cA
h
1/2
.
(2)Taking into a ount that
A
h
is an in reasing fun tion of time ( f. Murgia 1996), and assumingthatthedensitydistributionoftheunperturbed ambientgassurroundingtheradio sour eshas a power-law radialdensity distribution(the simpliedKing's (1972)
prole)s aling with the distan e r from the enter of the host galaxy as:
ρ
a
(r) = ρ
0
r
a
0
−
β
,
(3)where
ρ
0
is entral density of the radio ore,a
0
is its radius, andβ
is exponent of the power-lawdensityprole inthe King'smodel. Integration of Eg.(2)givesthe totalr
j
(t) = c
1
Q
jet
ρ
0
a
β
0
1/5−β
t
3/5−β
.
(4)where
r
j
is identied with one half of the sour e's linear size,r
j
= D/2
. A value of the oe ientc
1
is a omplex fun tion of equations of state of the unperturbed externalmediumand sho ked jetand surroundingenvironmentmaterial( f. Kaiser&Alexander 1997). The sho ked jet material inates a spa e volume (hereafter referred
to as o oon) within, so- alled, onta t dis ontinuity. The observed lobes of an FR II
type sour e are identied with the radiating parts of this o oon ( ross-hat hed area
inFigure (3)).
The lobeexpand due to the hotspotplasma pressure,
p
h
, and the o oonpressure itself-p
c
.Thepressure ratioisproportionaltothe axialratio-R
t
, denedasthe ratio of lobe's length toits base diameter:P
hc
≡
p
h
p
c
≃ 4 R
t
.
(5)Inparti ular,the oe ient
c
1
isdependentonP
hc
also( f.Eqs.(32) and(38) inKA). It is worth noti e that in the KA model the rate at whi h parti les are transportedfromthe AGN to the hot spotsis onstant during the lifetimeof the sour e. Iftwo of
itsbasi freeparameters:
Q
jet
andρ
a
(r) = f (ρ
0
, β)
,arespe ied,this simpliedmodel predi ts evolution of sour e's geometri alparameters only.3.2 Sour e energeti s and luminosity evolution
Inordertoprovideanother,independentrelationbetweentheparameters
Q
jet
andρ
0
, Kaiser,Dennett-Thorpe & Alexander(1997) ombinedthe pure dynami alKAmodelwith the analyti al model (hereafter KDA model) for the total energeti s of a FR II
type sour e, i.e.forexpe tedradio emissionof its lobes(or o oon) under inuen e of
theenergylosspro esses:adiabati lossesduetotheirexpansion,syn hrotronemission
inthemagneti eld,andinverseComptons atteringoftheCMBRphotons.Theexa t
formulafor the radio power is not an analyti ally solvable integral inthis model, but
it an be solved numeri ally for assumed values of the modelfree parameters, tra ing
allof the sour e's energy losses.
The initialenergy distribution of inje ted parti lesis given by
n(γ
i
) = n
0
γ
−
p
. The
additionalmodel free parameteres that have to be xed are:
a
0
,β
,θ
(the in lination angleofthejetaxistotheobserver'sdire tionofsight),γ
min
andγ
max
(Lorentzfa tors determining the energy range of the relativisti parti les),p
(initial power-law expo-nent),Γ
j
,Γ
a
,Γ
B
andΓ
c
(adiabati indi esintheequationof stateforthejetmaterial,the ratio ofthe energy density of the magneti eld andrelativisti parti les, given by
r = u
B
/u
e
= (1 + p)/4(1 + k
′
)
, assuming energy equipartition ondition.k
′
is a ratio
of thermalto relativisti parti les, and
p = 2α
inj
+ 1
.Anensemble of
n(γ)
relativisti ele trons withLorentzfa torγ
,pla edinavolume V,inthepresen e ofmagneti eldB
,emitssyn hrotronpowerperunitfrequen y and unit solid angle a ording tothe relation:P
ν
=
σ
T
c
6π
B
2
2µ
0
γ
3
ν
n(γ)V,
(6)where
σ
T
isthe Thomson ross-se tionandµ
0
isthe permeability of freespa e. These relativisti ele trons are supposed tobeinje ted into the lobe fromthe hotspot by itshead (extended regionof turbulent a eleration around the hotspot.) Radio spe trum
evolves with time due to energy losses. The pro ess of slowing down the parti lesthe
Lorentz fa tor
γ
an be des ribed by the formula:dγ
dt
= −
a
1
3
γ
t
−
4
3
σ
T
m
e
c
γ
2
(u
b
+ u
c
),
(7)wheretherstterm(fromtheright)referstotheadiabati lossesintheexpandinglobes
and the se ond - to the ombined syn hrotron and inverse Compton ee t radiation.
Inthis formula,
m
e
is the ele tron mass,u
b
- energy density of the magneti eld, andu
c
- energy density of CMBR.IntegratingEq.(7)overtimeandperformingsomeothertransformationsone nally
obtainsthe formulaforthe radiopowerof innitesimalvolumeelementsofthe o oon
atgiven frequen y. Bysumming the total ontributionof allof those elements we an
numeri ally al ulatethis integralovertheinje tiontime
t
i
.However,t
i
annotbeless than a minimum inje tion time,t
min
, at whi h the o oon material(still radiatingat frequen yν
) hasγ
i
≤ γ
max
. Thus:P
ν
(t) =
t
Z
t
min
dt
i
σ
T
cr
6πν(r + 1)
Q
jet
n
0
(4R
2
t
)
(1−Γ
c
)/Γ
c
×
γ
3−p
t
a
1
/3(p−2)
i
[t
−
a
1
/3
− a
2
(t, t
i
)γ)]
2−p
·
t
t
i
−
a
1
(1/3+Γ
B
)
.
(8)Asithavebeenmentioned,thereare examplesofmoreadvan eddynami almodels
basedontheoriginalKAmodel.Theyareallrelatedtothe samebasi onsiderationof
sour e'slengthvs.agedependen eandtheme hanismofradiationandradiativelosses.
energydistributiondenedabove.Inthissimplieds enario,theradiatingparti lesare
inje ted from the hotspot into the lobe. Some authors argue that this approximation
maybein orre t.Forexample,intheBRW(Blundell,Rawlings&Willot1999)model,
this inje tion index is not expe ted to be onstant and varies between the dierent
energyregimesduetosomebreak frequen ies.Onthe ontrary,theMK(Manolakou&
Kirk2002)assumesthatthisindexisalso onstant,buttheparti le'sbehaviourismore
ompli atedthanina aseof pureKDA modeland theyare additionallyrea elerated
in the lobe's head. The both models are widely des ribed and ompared by Barai &
Wiita(2006).
3.3 Determining dynami al age and other physi al parameters
for real FRII type radio sour es
The originalKDA modelallows predi tion of the jets length (i.e. the linear size of a
sour e) and its radio power at a given frequen y. This is possible if the values of the
allfreeparametersof themodelare spe ied.However, aspe ialtoolhas been desired
tosolve the "reverse problem", i.e.to estimate model's physi al parameters for a real
radio sour e with all the observational parameters derived from radio maps: the size
and axial ratio of its lobes and their radio spe trum, i.e. the luminosity at a number
of observingfrequen ies.
In order to solve su h "reverse problem" Ma halski, Chy»y, Stawarz & Kozieª
(2007a; hereafter referred to as MCSK) elaborated the algorithm "DYNAGE" whi h
allows toderivevaluesof fourof the sour e's unknown parameters (i.e.age, jet power,
entral density and the initial inje ted energy) from the t to the known observables
(sour e's size, volume, the radio power at given frequen y and the shape of the
spe -trum).This numeri al approa hdemands multifrequen y radio observations in luding
atleast three dierent ux densitiesembra ing possibly wide range of the radio
spe -trum. The "DYNAGE" algorithm demands also xing values of the remaining free
modelparameters, i.e. values of
a
0
,β
,θ
,γ
min
,γ
max
,Γ
j
,Γ
a
,Γ
B
,Γ
c
, andk
′
indi ating
ratioof energydensityof the thermalparti lestothat ofthe relativisti parti les.
Set-ting all of these parameters, one an estimate values of
t
,Q
jet
,ρ
0
, andα
inj
for every individualFR II type radio sour e.Espe ially,the dynami alage ofagivensour e anbe estimated.The "DYNAGE"
algorithm provides the re ipe for tting "the best" model radio spe trum to the
ob-served data. This pro edure is shown in details in MCSK. So- alled "age solution" is
found knowing that numeri ally tted values of the
Q
jet
in rease with de reasing age and, on the ontrary,ρ
0
tend to in rease with age ontheQ
jet
ρ
0
diagram ( f. their gure (1)). Therefore the rossing point for all of the frequen y urves on thesedia-grams(or, thenumeri allyted age orrespondingtotheminimumdispersionof these
dependssigni antlyon
α
inj
value.Thus,the"DYNAGE"algorithmpredi tsthe"best age solution"resultingfromthet tothe best "ee tive"value of the initialenergy ofinje ted parti lesprovidingaminimum of the jet's kineti energy,
Q
jet
· t
3.4 Limitations of the KDA model
The KDA model suers three serious limitations in its des riptions of the physi al
pro essesand onditionsgoverningthedynami alevolutionofradiosour es.Theseare:
1. Negligen e of observed asymmetries in the lobe's length, axial ratio, and radio
spe trum,
2. Unlimitedde rease of the externaldensity prole,
3. Constant relativisti parti les energy distribution and the jet power during the
lifetimeof asour e.
3.4.1 Asymmetry of radio sour es
A ording to the lassi al approa hbased on standard model of FR II type sour es,
every radio sour e forwhi h modelingis performed isstri tly symmetri al. Therefore,
the KDA model assumes that the two lobes have the same length, axial ratio, and
total radio power at a given frequen y. In the ase of real radio sour es more or less
evident asymmetries inboth: lobe's length and theirradio luminosity are observed. In
the standard modelingthese dieren es are negle ted and the values of
D
andP
ν
,the opposite for lobes, are averaged. This simpli ation implies that estimated values oft
,Q
jet
andρ
0
, as well asβ
exponent, are always equal for both lobes of the sour e. In the Se tion 4, I analyse dependen es of the ratios of the lobe size,D
1
/D
2
, and mono hromati power,P
1
/P
2
,resulting fromthe KDA model, ondierentvalues of a fewof itsbasi free parametersassumed for opposite lobes of a du ialsour e.3.4.2 Unlimited external density prole
The original KDA model assumesthat radio galaxyevolves in one-medium
environ-ment: ahalo with de reasing density des ribed by the simpliedKing's (1972)prole.
Gopal-Krishna& Wiita(1987) proposed a more sophisti ated modelof the jet
propa-gation inwhi h host galaxy issurrounded by two-media environment onsisting of an
X-ray halo aroundthe parent galaxy with gas density de reasing with radial distan e
from the galaxy and a hot IGM with onstant density. In Se tion 5, I revised their
modelintrodu ing ontemporaryvaluesforthe density and temperatureofthe
onsid-ered media, as well as ompare the revised model dynami alpredi tions with the age
andlinearsize ofseveral FR IItypesour es(formingthreedierentsamples)resulting
KDA model has been also onstru ted on the assumption that the jets a tivity
of the radio sour e is ontinuous, the power-law distribution of inje ted relativisti
parti les is onstant, and the lobes are powered with onstant amount of the new
radiatingparti les. This approa h is a good approximation only in the ase of young
radio sour es. It alsoimplies that originalKDA modelis appli able for young sour es
with regular spe tra only. However, the observations indi ate that there are many
examplesof radio sour es (in luding giant radio galaxies)in whi h the high-frequen y
parts of their spe tra are steeper than expe ted for the ase of the KDA
ontinuum-inje tion model,
α
inj
+ 0.5
. This strongly suggests that their a tivity stopped some time ago. In su h the ase, the use of "DYNAGE" algorithm results a di ulty indeterminingasatisfa tory agesolutionfor su hasour e by the best tofmodel'sfree
parameters to the observables. The attempt of extending this algorithm to a version
appli ableforFRIItyperadiosour eswithhigh-frequen yspe trasteeperthan
α
inj
+
4 Asymmetries of FR II type radio sour es predi ted
with the KDA model
4.1 Lobe's length and total luminosity asymmetries
As numerous observations indi ate, the asymmetries of FR II type radio sour e's
lobes (their morphology, size, brightness distribution, total luminosity, spe trum and
polarization of radio emission an be meaningful. In this dissertation I onsider the
lobes' length,
D
,and mono hromati radio powers,P
ν
, only.In most ases these observed asymmetries annotbe explainedby asimple
proje -tion ee t only- the situationwhen one of the lobes, that isphysi ally situated loser
to the observer (when the jet's axis is signi antly in lined to the dire tion of view)
seems to be longer and fainter than the opposite one, a ording to the dieren es in
the lighttravel time (Longair& Riley, 1979).
ThereforeinthisSe tion,Ianalyseasymmetriesbetweenthelobe'slengthandradio
luminosity predi ted by the KDA model. They appear in the model when the values
of some of it's free parameters are dierent for the opposite lobes of a given sour e.
Hereafter these model-predi ted asymmetries are des ribed by the ratios
D
1
/D
2
andP
ν1
/P
ν2
(the latter at the observingfrequen y of 178 MHz.Eqs.(4)and(8)inSe tions 3.1and3.2indi atethemodel'sparametersfromwhi h
the lobe's length,
D/2
, and its radio luminosity,P
ν
depends. Be ause the parametersQ
jet
,ρ
0
,a
0
andt
areexpe tedtoberather onstantforthegivensour e-relativevalues ofD/2
willbe onlydependent onthe oe ientc
1
andβ
parameter, respe tively. As it had been mentioned before, oe ientc
1
is a fun tion of the following model's parameters:Γ
a
,Γ
c
,β
andP
hc
≃ 4 R
2
t
( f. Eq. 5). However, one an assume that only values ofβ
andR
t
an be dierent in the opposite lobes. Similarly, relative values ofP
ν
depend onfollowingparameters:Γ
B
,Γ
c
,R
t
,β
,α
inj
andr = f (k
′
)
( f.Se tion 3.2), wherethe lastfour values onlyare supposed tobe dierent inthe opposite lobes. It isworth notingthat
R
t
value parametrizesthe assumed ylindri al geometryof the lobe ( o oon) and is used to al ulateits volume,V
.The remaining values of the KDA model's free parameters have been set up as
follows:
a
0
=
10 kp ,γ
min
= 1
,γ
max
= 10
7
,
θ = 90
◦
,
Γ
j
=Γ
B
=Γ
c
=Γ
a
= 5/3
.4.2 Asymmetries resulting from dierent
β
exponentOne an expe t that
β
exponentinthe power-lawdensity distributionofthe external mediumsurrounding the radio sour e may have dierent values in dire tions ofprop-dieren es will ause dierent values of both: lobe's size and its radio luminosity at
a given frequen y without referring to any internal asymmetries due to physi al
pro- esses inthe sour e itself.In orderto analysethe above asymmetries, I used adu ial
sour e with xed values of
Q
jet
= 10
38
W,ρ
0
= 10
−
22
kg/m3
,α
inj
= 0.51
(implyingp = 2.02
),R
t
= 3.0
andz = 0.5
.The al ulationisperformedfortwosele ted sour e'sages:
t =
10and100 Myr(forD
1
/D
2
asymmetry) and for three s enarios inthe ase ofP
ν,1
/P
ν,2
asymmetry:t = 10
Myr andk
′
= 0
,t = 100
Myr andk
′
= 0
, andt = 100
Myr andk
′
= 10
. Following Kaiseretal.(1997)I usedthe frequen y of 178MHz. Itisassumed that the valueofβ
may varyfrom1.0to 1.9, withthe highestdieren eβ
1
− β
2
= 0.45
between the lobes of this sour e.Figure (5) presents predi ted ratios of
D
1
/D
2
(plot a) andP
ν,1
/P
ν,2
(in log s ale, plot b), respe tively, versus the dieren eβ
1
− β
2
. The urves on both plots indi ate median values of the above ratios in a set of the model solutions forD
1
andD
2
, as wellas forP
ν,1
andP
ν,2
that are dependent on varying values ofβ
providingthe same dieren e ofβ
1
− β
2
. The verti al bars show the standard deviation fromthe median value.The diagramsshow that the lengthand luminosity asymmetries in rease with age
of the sour e. Besides, the asymmetry inluminosity in reases with in reasing fra tion
ofnon-relativisti (thermal) parti lesintheradiolobes.Open ir lesonboth diagrams
marktheobserved ratios of
D
1
/D
2
andP
ν,1
/P
ν,2
versusdieren e ofβ
1
− β
2
published forthesampleof30giant-sizedradiogalaxiesby sampleof30giant-sizedradiogalaxiesbyMa halskietal.(2009)andMa halski(2011).Intheabovepapers,
P
ν,i
(i = 1, 2)
are given at a number of observing frequen y of 151 MHzas the losest one to the modelfrequen y of 178 MHz. For the sample sour es (or lobes) withoutdata at 151 MHz, I
use the radio power al ulated with ux densities interpolated between neighbouring
data points, e.g. between 74MHz and 325 MHz. The values of
β
are taken from their "self onsistent" solution for the opposite lobes, i.e. from the DYNAGE t of theirradio spe tra with a model assuming ommon values of its free parameters
Q
jet
andρ
0
.A omparison between the modelpredi tion and the above data for real sour es is
0,0 0,2 0,4 0,6 0,60 0,80 1,00 1,20 1,40 1,60 t = 10 Myr t = 100 Myr D D 1 - 2 0,0 0,2 0,4 0,6 -1,00 -0,80 -0,60 -0,40 -0,20 0,00 0,20 0,40 0,60 t = 10 Myr, k' = 0 t = 100 Myr, k' = 0 t = 100 Myr, k' = 10 l o g P l o g P 1 - 2 b
Figure5:
D
1
/D
2
(plot a)andlogP
ν1
− logP
ν2
(plotb)vs.varyingdieren eβ
1
− β
2
forthe xedvalueof:α
inj
,r
,R
t
andtwodierentvaluesofdynami alagesandk
′
.The urvesindi ate
mediansofthe distributionsof allindividualmodelsolutions for
D
1
/D
2
andP
ν,1
/P
ν,2
ratios, al ulated asthefun tionof in reasing dieren esbetween theβ
parameters in theopposite lobes. The error bars are the standard deviation in these distributions. Cir les represent4.3 Asymmetries resulting from dierent axial ratio
R
t
Observations shows that the opposite lobesmay have signi antly dierent values of
R
t
. Following Kaiser (2000), it is asummed that the axial and transversal expansion of the radio lobe, governed by the presssure ratioP
hc
, is related to its axial ratioR
t
( f. Eq. 5). Therefore, observed dieren es inR
t
values in the opposite lobes suggest dierent ratios of the pressure inside the external environment along and a ross thejets (dierent values of
P
hc
at the ends of the opposite jets of the same sour e), thus dierent physi al onditions in the IGM surrounding the jet's material in luding itsinhomogenity.
Dierent values of
R
t
inuen es both the lobe's length and luminosity, besides dierentβ
exponent an ause an observed asymmetry of these lobe's parameters. I assume thatR
t
varies from 1.5 to 6 with the maximum ratio of 2. Figure (6) shows expe ted ratios ofD
1
/D
2
(plot (a) andP
ν,1
/P
ν,2
(in log s ale, plot (b), respe tively, versus theratioofdieren eofR
t,1
/R
t,2
al ulatedfortwodierentagesofthedu ial sour es -t = 10
andt = 100
Myr. Additionaly, plot b shows the model predi tion fort = 100
Myr andk
′
= 10
.Open ir les markthesame observationaldata asinSe tion 4.2.0,8 1,0 1,2 1,4 1,6 1,8 2,0 2,2 2,4 2,6 2,8 0,60 0,80 1,00 1,20 1,40 1,60 t = 10 Myr t = 100 Myr D D R T1 / R T2 0,8 1,0 1,2 1,4 1,6 1,8 2,0 2,2 2,4 2,6 2,8 -1,00 -0,80 -0,60 -0,40 -0,20 0,00 0,20 0,40 0,60 t = 10 Myr, k' = 0 t = 100 Myr, k' = 0 t = 100 Myr, k' = 10 l o g P l o g P R T1 / R T2 b
Figure6:
D
1
/D
2
(plota)andlogP
ν,1
−logP
ν,2
(plotb)vs.varyingratioofR
t
parameterfora xedvaluesof:α
inj
,r
,β
andtwodierentvaluesofdynami alagesandk
′
.The urvesindi ate
medians of the distributions of all individual modelsolutions (as in Se tion 4.2.) al ulated
asthefun tion ofin reasing dieren esbetween the
β
parameters intheoppositelobes. The error bars are the standard deviation in these distributions. Cir les represent observational4.4 Asymmetries resulting from dierent
α
inj
parameterIn ontrast to the parameters
β
andR
t
,α
inj
ae ts onlythe value of the mono hro-mati radio power at a given frequen y. It does not have any inuen e for the sizeof the lobe. Diagrams in this se tion show how a dieren e in the "ee tive" density
distribution of relativisti parti les a ross the bow-sho k at the head of the opposite
jets an result in a diverse syn hrotron emission of the relevant lobes and expe ted
symmtetryof
P
ν,1
/P
ν,2
.The al ulationsare performedwithidenti alvaluesofthefree parameters of the modelas inprevious twose tions, but hanging values ofα
inj
only. It is assumed that the values ofα
inj
may hange from 0.5 to 0.8, with the maximum dieren eα
inj,1
− α
inj,2
= 0.15
for the opposite lobesof the du ial sour e.Resulting diagram is presented in Figure (7). As previously, the urves show the
median values of
P
ν,1
/P
ν,2
in a set of the model solutions of this ratio for a given dieren e ofα
inj,1
− α
inj,2
. Again, the verti al bars show the standard deviation from these median values. Open ir les markthe observational data as in Se tions 4.2 and-0,05 0,00 0,05 0,10 0,15 0,20 -1,00 -0,80 -0,60 -0,40 -0,20 0,00 0,20 0,40 t = 10 Myr, k' = 0 t = 100 Myr, k' = 0 t = 100 Myr, k' = 10 l o g P l o g P inj1 - inj2
Figure7:
logP
ν,1
− logP
ν,2
vs.varyingdieren eofα
inj,1
− α
inj,2
parameterforaxedvalues of::β
,r
,R
t
andfordierentvaluesofdynami alagesandk
′
.The urvesindi atemediansof
thedistributionsofallindividualmodelsolutions(asinSe tion4.2.) al ulatedasthefun tion
ofin reasingdieren esbetweenthe
α
inj
parametersintheoppositelobes.Theerrorbarsare the standard deviation in these distributions. Cir les represent observational data available4.5 Asymmetries resulting from dierent
k
′
parameter
Dierentvaluesof
k
′
(theratioofenergydensityofthethermalparti lestothatofthe
relativisti parti lesinthelobe'smaterial)also auseanasymmetryinradioluminosity
of the opposite lobes. It ae ts only the value of radio power (Se tion 3.2) and has no
inuen e intolinear size of the lobe. Anestimation of the value of
k
′
inreal sour esis
verydi ult.Forexample,Bro ksoppetal.(2011),modellingthe doubledoubleradio
galaxyB1450
+
333, onsideredk
′
values from0 to100.
The al ulations are performed with identi al values of the free parameters of the
model as in previous two se tions, but hanging values of
k
′
only. It is assumed that
they may hange from0 ( orrespondingto the ase of nothermal parti lesin the jet)
to30,with themaximumratio
k
′
1
/k
′
2
= 15
for theoppositelobesof thedu ialsour e. Resulting diagram is presented in Figure (8). The urves show the median values ofP
ν,1
/P
ν,2
ina set of the modelsolutionsof this ratio for a given ratiosofk
′
1
/k
′
2
.0 2 4 6 8 10 12 14 16 18 -1,00 -0,80 -0,60 -0,40 -0,20 l o g P l o g P t = 10 Myr t = 100 Myr k 1 ' - k 2 '
Figure8:
logP
ν,1
− logP
ν,2
vs.varyingratioofk
′
parameterforaxedvaluesof::
β
,r
,R
t
and for two dierent values of dynami al ages. The urves indi ate medians of the distributionsof all individual model solutions (asin Se tion 4.2.) al ulated asthe fun tion of in reasing
dieren esbetween the
k
′
parameters in theopposite lobes. Theerror bars arethestandard
Observed dieren es in linearsizes and mono hromati radiopowers of the opposite
lobes of FR II type radio sour es are often too large to be explained by the sele tion
ee ts or proje tion of the sour e on the plane of the sky only. The al ulations
per-formed in Se tions 4.2, 4.3, 4.4 and 4.5 showed that su h asymmetries an be easily
predi ted with the KDA model assuming dierent physi al onditions in these lobes
and/ortheir environment.
These al ulations pre ise the inuen e of dierent values of
β
,R
t
andt
model parametersonthe lobe'slengthasymmetryD
1
/D
2
(asexpe tedfromEg.4),aswellas onrmasuppositionthatasigni antinuen eontheluminosityasymmetry,P
ν1
/P
ν2
, have themodel parametersβ
,R
t
,α
inj
,k
′
and
t
.Althoughdierentage ofthe opposite lobesisrather not admissible,this is worth emphasizingthat the inuen e ofabsoluteage value of the du ialsour e onthe asymmetries onsideredhere ismeaningful.
4.6.1 Model predi tions for the asymmetries
Asymmetryoflobe's lengthdepends onthe valuesofmodelparameters
β
andR
t
.This inuen eis stronger inthe ase ofR
t
parameterdeterminingthe pressure ratiosinthe lobe,P
hc
( f.Figure(6a)).In turn,Figure(5a)presents very strongdependen yofthe sour e's age onthe resultingD
1
/D
2
asymmetry. These asymmetries in rease with the age of a given sour e.Asymmetries of lobe's radio brightness may result from dierent values number
of the model parameters. In Se tions 4.2, 4.3, 4.4 and 4.5 I analyzed the dependen e
of su h asymmetry on the values of parameters
β
,α
inj
,R
t
andk
′
. A variation of
the
β
parameter gives relatively lowP
ν,1
/P
ν,2
asymmetries slowly in reasing with the in reasing age of the sour e,while for parametersα
inj
,R
t
andk
′
the predi ted degree
of the asymmetry isnearly the same and highlydependent onthe sour e's age.
As on erns the
P
ν,1
/P
ν,2
asymmetry, one an ompare its dependen e on the pa-rametersβ
,α
inj
andR
t
with that resulting from dierent values ofk
′
. It is almost
equally strong as the impa t of the hange of the sour e's age on shape of the model
urves. At the same time it isdi ultto learly determine whi h one of these
param-eter inuen es the greatest asymmetry. In the ase of the parameters
β
andα
inj
it is the age, but in turn forR
t
it isa hange of the parameterk
′
.
It is learly seen that whilein the ase ofasymmetry inthe lengthof the lobesthe
number of fa tors that may ause them is relativelysmall, whereas the asymmetry of
their radio brightness may be aused by dierent values of several model parameters
and thusit isnot easy todetermine whi hone of them has a de isiveinuen e inthis
Asymmetries of lobe's length and brightness predi ted by the model are onfronted
with the observed asymmetries in the smalland heterogeneous (only available in the
literature)sampleof30"giant"radiogalaxies(Ma halskietal.2009,Ma halski2011).
In both of these publi ations the observed asymmetries
D
1
/D
2
andP
ν,1
/P
ν,2
are re-produ ed by varying the values of only two model parameters, namelyβ
andα
inj
. In parti ular,dierentvaluesofβ
inthe oppositelobes(where usually dierent values ofR
t
are known from observations) are used to explain the observed asymmetry in the lengthof the lobes. In the next step, the observed asymmetry in the brightness of theoppositelobesof radio galaxiesfromthose samplesisreprodu edby a variationofthe
α
inj
parameter in these lobes. In this way the observed ratio ofP
ν,1
/P
ν,2
is des ribed by a diversity of not the only one, but three parameters of the model,R
t
,β
andα
inj
. Thus these observations onrm the on lusions from the previous paragraph that agivenasymmetryoflobe'slengthandbrightness annotbedes ribedbydierentvalues
of a single modelparameter, however it an be provided by a ombination of several
dierent values of its free parameters only. This is in a ordan e with the physi s of
theFR IItyperadiosour e,be ausetheradio brightnessatagiven frequen y depends
onenergydistributionofrelativisti parti les, strengthandorientationofthe magneti
eld, and the volume of the sour e (or its lobes), thus on mu h larger number of the
model'sfree parametersthan the lengthof the lobes. Itis alsoevident fromthe
exam-pleof the observationaldataof thesamplesour es. The diagramsillustrating
P
ν,1
/P
ν,2
asymmetry show mu h more noti eable dispersion of the sample sour es around themodel urvesthan the diagrams onfronting the observed
D
1
/D
2
asymmetry withthe modelpredi tions.Onlythesmallfra tionofobservedasymmetries orrespondstothepredi tionofthe
modelpresented onFigures(5), (6)and(7). Itis worthnoting thattheseasymmetries
in real radio sour es mainly orrespond to the model urves al ulated for sour e's
age of100 Myr and for dierent values of
k
′
. Itis understandablebe ausethe samples
omprisesour eswithverylargelinearsizesandagesofabout
≃
100Myr,notin luding, however, smalland relativelyyoung sour es.Inthe aseofthediagramspresentingthe
P
ν,1
/P
ν,2
asymmetryone annoti estrong deviationsofthenotablepartoftheobservedsour esfromthemodelpredi tions.Thesesour es are not even lose to the urves al ulated for the sour e age of 100 Myr (for
both values of
k
′
), though their tted dynami al ages al ulated for the sour es have
typi alvalues of this range and are not ex eeding the value of 250 Myr.
Theabove on lusionsarenot omprehensiveinthesensethatthesampleisnot
rep-resentative for the entire FR II typesour es population. As itwas alreadymentioned,
thesele tionee t ausesthatit onsistoflargeradiogalaxiesonly.Therefore,afuture
resear h shouldfo us onthe omparisonof modelpredi tions for the asymmetry with
radio-jets propagation through the X-ray haloIGM
interfa e (Kuligowska et al. 2009)
Kuligowska, Jamrozy, KozieªWierzbowska & Ma halski,
2009, A A, 59, 431
Extended large-sized radio sour es are not easy to re ognize be ause of their
relatively low radio brightness and a di ulty to dete t eventual bridge onne ting
brighter parts (lobes) of a ommonradio stru ture. Several observational eorts show
that most of known GRGs lie at low redshifts of z
<
0.25. For a long time this aused a presumption that su h extragala ti double radio sour es, espe iallythose ofFRII-type, did not exist at redshifts higher than about one be ause of the expe ted strong
evolutionofauniformIGM,
ρ
IGM
∝ (1+z)
3
, onningthelobesofsour es(e.g.Kapahi
1989). The situation hanged over 10 years ago when Cotter, Rawlings & Saunders
(1996)and Cotter(1998) presented anunbiasedsample ofgiantradio sour essele ted
from the 7C survey (M Gil hrist et al. 1990). Their sample omprised 12 large-size
sour es with 0.3
<
z<
0.9. The list of known GRGs with z>
0.5 and D>
1 Mp is very short.The undertaken sear hforsu h GRGsonthe southern sky hemispherewiththe11m SALT teles opeduring the Performan e Veri ation (PV) phase has resulted in
the dete tion of 21GRGswith the proje ted linear size greaterthan 1Mp .However,
one anfoundthattheirredshiftsdonotex eedthevalueof0.4andtheenergydensity
in only two of them is less than
10
−
14
Jm
−
3
. One of them, J1420-0545, is the largest
known GRGin the Universe ( f.Ma halski etal. 2008).
Thedynami alevolutionofaFRIIradio sour estronglydependson hara teristi s
of the ambient medium. Gopal-Krishna & Wiita (1987) proposed the two-medium
model onsistingofanX-rayhaloaroundtheparentgalaxywithgasdensityde reasing
with radial distan e from the galaxy and a mu h hotter intergala ti medium (IGM)
with onstantdensity.Thesetwomediawere on eivedtobepressure-mat hedattheir
interfa e.Their modelallowed topredi tlimiting(maxima)values forthesour e's age
andlinearsizedependingontheenvironment onditions,thejetpower, andthe osmi
epo h hara terized by the sour e's redshift. However, our re ent dete tions of very
large-sized radio sour es with z
>
1 and ex eeding the limits predi ted by their model (hereafterreferredtoasG-KWmodel),suggeststhatsomeofitsfreeparametersshouldbe modied.
In this Se tion an observational onstraint for the G-KW model is analyzed. For
this purpose, aneort todetermine the highestsizes anddynami alages of FRII-type
radio sour es at redshifts 1
<
z<
2 is undertaken. The original G-KW model is briey des ribed and modied adopting modern ( ontemporary) values for thermodynamihead and the age are al ulated.
Theobservationaldatausedto onstrainthetwo-mediummodelarepresented. The
smallsampleof themost distantgiant-sizedradio sour esisrevisedand supplemented
withtwootherlimitedsamplesofFRII-typesour es omprising:(i)sour eslargerthan
400 kp within the redshift range 1
<
z<
2, most of them found in this paper, and (ii) sele ted3CRRsour esinmajoritysmallerthan400kp atz>
0.5forminga omparison sampleof "normal"-sizedradiosour es. Physi alparametersofthe samplesour es:thedynami al age, the jet power, the entral radio- ore density and the IGM density,
and others, are derived using the "DYNAGE" algorithm (Ma halski et al. 2007a).
Theappli ation ofthis algorithmtothesamplesour es andthe resultingvalues ofthe
sour e'sparametersaredes ribedinSe tion5.2.A omparisonofthemodelpredi tions
with the observational data is presented and dis ussed inSe tion 5.3.
5.1 The base of the revised G-KW model
Inthe G-KW model,the jetpropagates intoatwo- omponentmedium omprised of:
thegaseoushalowithapower-lawdensityprole
ρ
h
(d) = ρ
0
[1 + (d/a
0
)
2
]
−
δ
bound
totheparentopti algalaxy,where
ρ
0
anda
0
arethedensityandtheradiusofthe entral radio ore, respe tively, andδ
=5/6. This distribution is assumed to be invariant with redshift. It is also assumed that this halo has nearly uniform ele tron temperature(kT )
h
[keV℄ (medium1),andthe surrounding hotter IGM of uniform density,
ρ
IGM
, with the temperature(kT )
IGM
(1 + z)
2
[keV℄ (medium2).SimilarilytoGopal-Krishna&Wiita(1987)itisne essery toassume hara teristi
values for the density and temperature of the onsidered media. The values adopted
hereafterfor the two omponentsare based on the followingdata:
(1)The radio oreradius,
a
0
=3kp is basedonthe tted X-ray surfa e-brightness proleofninenearby,low-luminosityradiogalaxiesre entlyobserved byCroston etal.(2008). This value of the radius is derived from the observed angular radius of about
10ar se .
(2) The halos' gas temperature have been determined in a number of papers. A
uniform temperature
(kT )
h
=0.7 keV was measured for a few nearby, X-ray luminous ellipti al galaxies with the Chandra Observatory by Allenet al. (2006). UsingXXM-Newtonand Chandra observations,the values from1to5keV with amedianofabout
2.1 keV was found by Belsole etal.(2007) for the X-ray lusters surrounding 20
lumi-nous 3CRR radio sour es. For the low-luminosity radio galaxies analysed by Croston
et al. (2008), a medianof the tted temperatures is about 1.4 keV. Taking the above
data intoa ount one an estimate:
(kT )
h
=1.4 keV. (3)The halos'gas(proton) density of (1 2)×10
4
m
−
3
istted to X-ray ountsby
pressure againstthepressure distributioninthehalo. Anon-relativisti gasinthermal
equilibriumthat hasanele tron density
n
e
[m−
3
℄and temperature
(kT )
e
[keV℄ willhave anele tron pressurep
e
=n
e
(kT )
e
[Pa℄. Expressingele tron density by the mass density,ρ
=n µ m
H
, this balan ewill have pla e atthe halo'sradiusR
h
al ulated fromρ
0
µ
h
m
H
1 + (R
h
/a
0
)
2
−
δ
(kT )
h
=
ρ
IGM
µ
IGM
m
H
(kT )
IGM
,
(9)where
µ
andm
H
are the mean mole ular weight and the mass of hydrogen atom, respe tively.µ
h
is assumed toµ
h
=0.5 andµ
IGM
=1.4. Besides, for the halo (medium 1)n
p
=1.5 × 10
4
m
−
3
is adopted (i.e. a mean proton density of the values given by
Belsole et al. (2007), whi h orresponds to
ρ
0
=10
−
22.6
kgm
−
3
, and the temperature
(kT )
h
=1.4 keV). Forthe IGM density 50% of the osmi matter density is taken, i.e.ρ
IGM
=0.5Ω
m
h
2
ρ
clos
=0.5 × 0.27 × 0.71
2
× (3 H
2
0
)/(4π G)
, whi h givesρ
IGM
=10
−
26.9
kgm
−
3
. For the IGM temperature the values of
(kT )
IGM
=25 keV is adopted. Substi-tuting the above values into Eq.(9) one an ndR
h
=642 kp . This radius of X-ray halo is ompatible with the radii determined by Cassano et al. (2007) for 15 Abellluster radio haloes with the mean of
∼ 560±
170 kp . This is worth to noti e that thit radiusof 642kp ismu hlargerthan 171 kp used by Gopal-Krishna&Wiita.Inanexpanding and uniform IGM this radius should evolveas
R
h
(z) = 642(1 + z)
−
5/(2δ)
kp , i.e.
642(1 + z)
−
3
kp for
δ
=5/6.Figure (9) (a), (b), ( ) present the basi hara teristi s of the two-media model:
the mass density
r(d)
, the ele tron temperaturekT (d)
, and the resulting ele tron gas pressurep(d)
, as fun tionsof the distan e from the host galaxy ( ompa t radio ore), respe tively. Note that the balan ep
h
(R
h
) = p
IGM
atd = R
h
orresponds to a rapid transitionbetweenr(R
h
)
andr
IGM
,aswellasbetweenkT (R
h
)
andkT
IGM
,and auses an unphysi al ee t shown in Se tion 5.6.1. The dashed urves in Figure (9) (a) and(b) indi ate desired smooth transitions between the relevant parameters whi h would
5.2 Predi tions of the model
For
d ≤ R
h
it is assumed that the jet propagate (through the medium 1) with a onstant opening angle,θ
. Under this ondition, the ram pressure balan e results in thefollowingdependen es forthe jetlength(theradiolobesize,D
)ontime(thelobe's age,t
) and the jet's head expansionvelo ity,v
h
, onD
ort
:D(t) = [(2 − δ) A t]
2−δ
1
,
(10)v
h
(D) = A D
(δ−1)
,
(11)v
h
(t) =
h
(2 − δ) A
δ−1
1
t
i
δ−1
2−δ
,
(12) whereA ≡
4 c
1
Q
jet
πθ
2
c ρ
0
a
2δ
0
1
2
.
(13)Here
c
1
isa onstantwithavaluebetween1.5and3.8dependingonthesour e's(lobe's) geometrydes ribedbyitsaxialratioR
t
(Kaiser&Alexander1997),whilec ≈ v
jet
isthe speed oflight.Thejet'sopening angleisalsodes ribed byR
t
,θ
2
= c
2
/(4 R
t
)
,wherec
2
isa onstantwithavaluebetween3.6to4.1dependingonspe i heatsforthematerialinthe jetand the lobe( o oon), (Eq.(17) inKaiser&Alexander 1997).At
d = R
h
(z)
the jet enters the hotter IGM (medium 2) at least an order of magnitude less densebutpressure-mat hed,asshown inFigure(9).Inordertoanalyzethe jet'spropagation
overthis regime,Gopal-Krishna& Wiitahave onsidered two likely extremes enarios
for the lobe's expansion:
S enario Awherethe jet openingangle,
θ
,is onserved. Due toarapidde rease of the ambient density at the interfa e,ρ
IGM
≪ ρ
h
(R
h
)
, a su ient ram-pressure will be provided only if the jet's head velo ity,v
hs
, in reases abruptly atd
=R
h
and then graduallyapproa hes thev
h
∝ d
−
1
lawexpe tedfora onstantdensitymedium.Inthis
s enario:
D(t) =
(
2
K(z) + a
δ
0
A
ρ
0
ρ
IGM
(1 + z)
3
1/2
t
!)
1/2
and
(14)v
h
(t) = a
δ
0
A
ρ
0
ρ
IGM
(1 + z)
3
1/2
/D(t),
(15) whereK(z) =
1
2
R
2
h
(z) − R
(2−δ)
h
(z)
a
δ
0
2 − δ
ρ
0
ρ
IGM
(1 + z)
3
1/2
(16)is a redshift-dependent onstant providing that the time orresponding to
D
=R
h
inEqs.(10) and (14) is the same.S enario B wherethe jet'sheadvelo ity a rossthe interfa e remains ontinuous
and thereforemat hed tothe value given by Eq.(11) for
D
=R
h
.This an be a hieved onlywith anabrupt aringof the jet's opening angle.Under this ondition the modelpredi ts: