Peter Mészáros,Pennsylvania State University

UHECRs,GRBs, HNe & Gal/Clus.

San Vito di Cadore. Dolomites, March 2014

GRB “Standard Model”

Mészáros grb-gen06

Shock formation• Collisionless shocks (rarefied gas)• “Internal” shock waves: where ? If two gas shells ejected with Δ Γ = Γ1-Γ2~ Γ, starting at time

intervals Δt ~tv ,, they collide at ris ,

ris ~ 2 c Δt Γ 2 ~ 2 c tv Γ 2 ~ 1O12 t-3 Γ 22 cm (internal shock) [Alternative picture: magnetic dissipation, reconnection]

• “External” shock: merged ejected shells coast out to res, where they have swept up enough enough external matter to slow down, E=(4p/3)res

3 next mp c2 Γ 2,

res~ (3E/4pnextmpc2)1/3 Γ-2/3 ~ 3.1O16(E51/nO)1/3 Γ2-2/3 cm

(external shock)

Schematic IS-ES (St.Mod.)Bulk Lorentz

factor

int. shockphotosphere ext. shock

int. shock

Internal shocks• Randomly

ejected shells of different Loretz factors

• Faster shells catch up with slower ones

• Internal shock randomizes relative kinetic energy

• Accelerate particles

External (Reverse+Forward) Shock

• Relat. ejecta (continuum of merged shells) decelerates due to swept up external metter

• Forward shock moves into ext. matter, reverse shock moves into ejecta

• Accelerate particles

Mészáros

Standard GRB shock leptonic EM rad’n:shock Fermi acc. of e- → synchrotron and inv.Compton

• GRB 990123 → bright (9th mag) prompt opt. transient (Akerlof etal 99) .

– 1st 10 min: decay steeper than forw.sh.• →Interpreted as reverse shock ....• Several more examples (but not ubiquitous)

­

νFν

ν

Sy (reverse)

Sy (forward)

0pt γ

Ext.Sh.

GeV

TeV

Or →?

E2N(E)

Mészáros, L’Aqu05

AG: Shock Sy/IC (SSC) Spectrum vs. time• Non-thermal power law

spectrum, both in int. and ext. shocks, due to

• Synchrotron, peak at ~200 keV (or ~ eV?)

• Inv. Compton, peak ~ GeV (or ~200 keV ?

• Sy peak location, ratio Sy/IC dep. on Bsh , ge,m

• Peak softens with time• Ratio Sy/IC decr w. time

E

E2NE

Sy

IC

t

tF

E

E

X,OR g

Also obvious alternative Sy/IC: EIC

Ackermann+14Sci 343:42

LAT-GBM Flat/Fgbm~0.2

GRB130427A bursting forth:

GBMLAT

Ackermann+14Sci 343:42

Ackermann+14Sci 343:42

2nd component!↘

Test case• GRB 130427 A is the best studied GRB so far

from radio to ~93 GeV; lots of information

• Various natural extensions of the (unreasonably simple) standard model can explain the observations so far

• But still: not enough observations and detailed calculations to decide which of the several possible extensions are at work in this particular GRB (sample of one -others: less data)

What aboutVHE neutrinos?

• At z=0.34 and Liso ~2x1054 erg, GRB130427A was expected to be best candidate so far

• But, IceCube : No Detection! (GCN 14520)

• Is this suprising?

• At least 2 reasons why not:

1. Expect no ν-detection in standard int. shock for this GRB (in IC40+59, one year of observation)

2. Expect no ν-detection in one out of two possible ‘non-standard’ models for this GRB (in IC40+59)

Standard(+) Model of GRB

Int. & ext. shocks,accelerate electronse,B →γ (leptonic);and accel. protons too (?)pγ→ν, γ (hadronic)

↙internal shocks

↙external shock

as UHECR/NU source

GRB VHE neutrinos

• Rd, Rph : dissipation radius where particles accelerated

• Fermi (or mag, reconn.) accel. dNp,e/dE ~ E-q

• e±,B →γ

• p,γ→π±→μ±,νμ→e±,νe,νμ

• For PL dNe,p/dE and dNγ/dE → dNν/dE also PL

• Parameters: ϵp , ϵe , ϵB : energy ratios of p,e,B to Etot

• Etot: total burst energy, Γ: bulk Lorentz factor

At the simplest level:

Simple p,γ→Δ→... model

17

pγ→Δ interaction of a Band-γ spec. with an E-2 proton spec. is

(WB 97)

γ

p

ν

Mészáros, NOW 04

νs from pγ from int. & ext. shocks

• Δ-res.: E’p E’γ ~0.3GeV2 in comoving frame, in lab:

→ Ep ≥ 3x106 Γ22 GeV

→ Eν ≥ 1.5x102 Γ22 TeV

• Internal shock pγMeV → ~100 TeV νs

• ( External shock pγUV → ~ 0.1-1 EeV νs )

• Diffuse flux: det. w. km3Waxman, Bahcall 97 & 99 PRL

Internal shock(prompt)

External shock

NOTE: internal shock (old paradigm ) + simple Δ-res. approx.

More generic (still simple)pγ neutron/neutrino sp:

• Proton sp. dN/dE ~E-q

• Photon sp. dN/dε ~ (εα , εβ ) for (ε<εb, ε>εb) with α~-1, β~-2.5,

• ph.sp.break: εb ~106ε6 eV → proton break:• Eb ~ 0.4 (Γ2/εb) GeV2 ~4x1016 (Γ22.5/ε6) eV• Neutron emission rate spectrum is

QCR(E)~Q0 (E/Eb)-q/[1+(E/Eb)β-α]• Neutrino energy is Eν ~ ηE , with η~0.06 • Neutrino emission rate (per flavor) is

Qν(Eν) ~(1/η)QCR(Eν/η)

IC40+59 search for VHE nus from 190 GRB (105 northern)

• Analyze 190 GRBs localized w. γ-rays betw, Tstart & Tend

• Use the WB’97 and Guetta’04 proton acceleration model in internal shocks, with E-2 spectr, εp/εe=10, and pγ→Δ-res →νµ

• Nu-flux normalized by obs. γ-ray flux, get Fν for 190 (right axis), and diffuse flux (all) assuming 677 yr-1

20

Nature 484:351 (2012), the Icecube collab.; Abbasi and 242 others (incl. P.M.)

(A) Model dependent search

Model

0.27

Data is factor 0.27 below model Fν at 90% CL

Model-independent ULs

• Take observed γ-ray flux and γ-ray spectral break energy

• Use these to infer the ν flux and the ν spectral break energy

• The ν-flux UL implies a UHECR proton flux UL, given by hatched region

• But (internal shock) models capable of explaining observed UHECR flux (data pts) are above this UL

IC59 2-year conclusions (190 GRBs)

• The fireball (more accurately: internal shock) model overpredicts the TeV-PeV nu-flux by a factor 3.7 ,(asuming Lp/Le=10, Δ-res only, Lorentz Γ~300-600)

• For a model independent fit, the 95% CL nu-flux UL is 2-3σ below what would be expected if the GRBs contribute most of GZK UHECR flux.

• In these models, either Lp/Le must be substantially below factor 3-10 assumed here, or the production efficiency of neutrinos is lower than was assumed.

22

Nature 484:351 (2012)

• Int. Shock model is expedient: best documented so far, and easy to calculate ⇒ main reason why its use is widespread

• But ... int. shocks known (past 10 years) to have difficulties for gamma-ray phenomenology (efficiency, spectrum, etc) - ongoing modifs. to Int.Sh. not yet well explored for CR/nus

• Also, acceleration rate of protons vs. electrons is unknown; are protons injected into accel. process at = or ≠ rate as electrons? Only energy restriction on model is Lp/Le ≲ 10.

• Other alternatives to internal shock: being investigated (are less easy to calculate). Progress, however, is being made.

• Even if GRB are not GZK sources, model indep. searches leave possibility of lower, observable nu-fluxes from GRB

23

Note that .....

Conclusion on the conclusions:

• These are conservatively stated conclusions• The first time VHE nus have put a significant

constraint on a well-calculable astrophysical UHECR-UHENU model, at 90-95% CL

• This is very valuable• Icecube is doing exciting astrophysics• A significant first step towards GZK physics

24

But, on closer look at IC3 analysis...

• IC22-59 analysis used a simplified version of WB97 , which results in overestimated model nu-fluxes

• Assumed FνIC /Fp = (1/8) fπ,b , where 1/8 because 1/2 pγ lead to π+ and each νµ carries 1/4 Eπ, and fπ,b=fπ (E=Eb)

• But fπ (E) = fπ,b is OK only for Eb < E <Eπ,cool; • Also for E<Eb , have fπ ∝E, because of decr. # of photons• Net result: model νµ flux overestimated by factor ~5• In addition, ignored multipion, Kaon decay, etc.

25

(Li ’12, PRD 85:027301)

Graphically....

• fC: energy of all photons approx. by break energy• f≈ : 0.7 (rounding error in one eq.)• fσ : 2/3(negelected width of Δ-res.)• Cs : correct en. loss of second’s & E-dep of proton mfp

26

Even using the (old paradigm) int. shockmodel, the simplifications have an effect

• IC-FC: IceCube Fball. Calc.

• RFC: “Revised” Fball.Calc. ⇒• NFC: Numerical Fball.Calc.

(Hümmer et al ’11, PRL 108:231101)

So for this model .... • Even when using old paradigm of internal shock, but more detailed physics (incl. multi-pion and Kaon production, ≠ cool’g. break for π, µ, and numerical as opposed to analytical calcul.)

• ⇒ Fν predictions below IC40+59 !

27

(← Hümmer et al, ’11, PRL 108:231101)

(Also: Li, ’12, PRD 85:027301 ;He, et al ’12, ApJ 752:29)

• Internal shock model use is expedient: best documented so far, and easy to calculate ⇒ reason why its use is widespread

• But ... int. shocks known (past 10 years) to have difficulties for gamma-ray phenomenology (efficiency, spectrum, etc)

• And, acceleration rate of protons vs. electrons is unknown; are protons injected into accel. process at = or ≠ rate as electrons? Only energy restriction on model is Lp/Le ≲ 10.

• Even if GRB are not GZK sources, model indep. searches leave possibility of lower, observable nu-fluxes from GRB

• Alternatives to int. shock: being investigated (less easy to calculate. Progress on photospheric & hadronic models

28

In summary .....

Back to GRB 130427A: nus? Gao, Kashiyama, Mészáros 2013, ApJ, 772:L4

Model independentInternal shock model

Calculate ULs using all recent corrections to the physics

130427A baryonic & magnetic photosph. modelGao, Kashiyama, Mészáros 2013, ApJ, 772:L4

In other words:

• IceCube neutrino ULs on the diffuse nu-bkg do not constrain the IS model so far; will need several more years to get near the ULs

• IceCube does not constrain the IS model’s ability to provide the 1019-1021 eV UHECR

• Even moderate L~1053 erg/s GRBs with fp~10 are able to explain GZK CRs (but below the ankle need other sources - always did)

• Also, they do not explain the diffuse PeV nus

Mészáros

Recent paradigm shifts inGRB prompt emission:

i) De-emphasize internal shocks (inefficent)

→ dissipative photospheric models

ii) Modify internal shocks : slow heating, (a) turbulence behind shocks (Fermi 2nd ord), (b) magnetic dissipation (high rad. efficiency), (c) hadronic cascades (naturally slower heat’g)

or:

Mészáros

GRB Paradigm shift (i)• OLD: internal + external shock (weak phot.)

• Photosphere: low rad. effic., wrong spectrum

• Internal sh.: good for variability, easy to model ; but poor radiative efficiency

• External sh.: was, and is, favored for afterglow model

• NEW: phot. + (int.sh? mag.diss?)+ext. shock

• Photosphere: if dissipative, → good rad. efficiency

• Int. sh: if magnetic, may be absent; but mag. dissip?

• External shock: most of GeV and soft afterglow

Paradigm shift(i): PhotosphereWhat are photospheric models?

& further Paradigm(i) sub-shift:

νµ from magn.dissip.phot. + ext.sh.

36

Gao, Asano, Mészáros ’12, JCAP 11:058

Compare mag.phot, bar.phot, int.shock @ var. η (or Γ)

37

(tvar=1 ms)

Gao, Asano, Mészáros ’12, JCAP 11:058

Diffuse nu from MPh, Bph, IS, η=300 & IC3 lim.

38

Gao, Asano, Mészáros ’12, JCAP 11:058

Diffuse nu from MPh, Bph, IS, η=1000 & IC3 lim.

39Gao, Asano, Mészáros ’12, JCAP 11:058

That is:

• IceCube neutrino ULs on the diffuse nu-bkg do not constrain the fixed radius, i.e. steady state IS model so far; will need several more years to get near the ULs

• IceCube ULs on GRB130427A do not constrain either the IS or baryonic phot. models, but do constrain mag. phot. model

• Will need consider better, time-dep models incl. effects of pair formation for model fits

Note that

• All previous GRB neutrino/CR calculations were time-independent (steady-state)

• Now can go one better on that

Time-dependent CR escape,gamma-rays & neutrinos

• Assume Internal Shock model but allow for shell motion and expansion, over R0 - 30R0.

• Use Nakar-Piran’02 variab. time distrib, and Wanderman-Piran’10 Lum. distrib., z<6

• Initial radius R0 dep. on δt; ph. sp. assumed Band, obeying Ep-Liso Yonetoku-Nava relatn.

• Inject protons E-2 spectrum, Γ=300, fp=10 as benchmark

• MC cascades, 2 cases: neutron conversion and sudden release

Asano & Mészáros, ApJ acc., arX:1402.6057

Spectra at source

• Time integrated spectra from one shell (at source)

• Shown for δt=0.1 s and various lumin, for theneutron conversion model

• Blue is νμ + νe and their anti-flavors, before oscillation

Asano & Mészáros, ApJ acc., arX:1402.6057

CR + nu Diffuse Bkg

• Diffuse CR (black) and νμ+anti-νμ (red, after osc) for neutron conversion model.

• Thin dash-dot: CR w/o photomeson & BetheH ; Dash(dot): CR/ν w/o GRB L >1054(1053.5)erg/s

• Thin red: cosmogenic nu

• Gray thick: IC3 40+59 diffuse UL for brokenPL

• Diamond: our integrated flux assuming broken PL & uncert. break range

Asano & Mészáros, ApJ acc., arX:1402.6057

CR + nu Diffuse Bkg

• Diffuse CR (black) and νμ+anti-νμ (red, after osc) for sudden release model.

• Thin dash-dot: CR w/o photomeson and BH ; Dash(dot): CR/ν w/o GRB L >1054(1053.5)erg/s

• Thin red: cosmogenic nu

• Gray thick: IC3 40+59 diffuse UL for brokenPL

• Diamond: our integrated flux assuming broken PL & uncert. break range

Asano & Mészáros, ApJ acc., arX:1402.6057

In other words:

• IceCube neutrino ULs on the diffuse nu-bkg do not constrain the standard IS model so far; will need several more years for that

• IceCube does not constrain the IS model’s ability to provide the 1019-1021 eV UHECR

• Even moderate L~1053 erg/s GRBs with fp~10 are able to explain GZK CRs (but below the ankle need other sources - always did)

• Also, they do not explain the diffuse PeV nus

Other UHECR sources?

• Hypernovae, a.k.a. jet-driven SNe, or transrelativistic SNe

• Star-burst galaxies (SBGs)

• Intergalactic cluster shocks (IGSs)

• AGNs (jetted? RQs?)

Hypernovae

48

← supernova SN 1006 (X-ray)- Hypernovae: similar but ~ 10-102 times more energetic; and portion of ejecta reaches ≥ semi-relativistic speed, possibly anisotropic ~500 times the rate density of GRBs

Hypernova ejecta as UHECR sources

• Type Ib/c but isotropic equiv EHN~3-5x1052 erg• 500 times GRB rate, and 10-1-10-2 usual SNIa rate• Semi-relativistic (v~c, or Γβ≥1) comp. in outflow

(shock accelerates down the envelope gradient) • Assume shock expands in WR progenitor wind,

magnetic field fraction εB of equipartition

49

Max. CR energy:

→ Proton: Emax ~1019 eV, and Fe: Emax ~2.6 1020 eV

(Wang et al, 2007, PRD 76:3009; Budnik et al, 2008, ApJ 673:928)

Energy distribution with velocity

Normal SN

Very steep distribution -> negligible contribution to high-energy CRs

Semi-relativistic hypernova: high velocity ejecta with significant energy

Berezhko & Volk 04

Wang, Razzaque, Meszaros, Dai 07 PRD 76:3009

Data from Soderberg et al. 06

CR spectrum:

HN as heavy UHECR source

• After 200 days, Fe, O survives photodissociat.

• HN from WR stars, preferred likely scenario QCHE → outer layers Fe rich

51Wang et al ‘08, ApJ 677:432

Astrophys. PeV nus: CR tracers?

IC3col, 2013 PRL 111.021103

28 Sub-PeV-PeV nus (<2013)

IC3 col, 2013, Sci., 342, 11

Any obvious sources???

Could these be HNe?

• A small fraction of PeV sources close to the Galactic Center might be the galactic TeV uni-ID sources - which in turn could be HNe; and CR protons of 10-100 PeV, via pp → PeV νs ; BUT: only 1/28 best fit - or 3/28 at 90% CL (Fox, Kashiyama & PM, 2013, ApJ, 774:74)

• Better guess: the predominantly isotropic component might be due to extragal. HNe, in ultraluminous IR galaxies- ULIRGs; or starburst galaxies - SBGs (He, et al, 2013, PRD, 87:063011)

More generally:

• Could pp sources might explain the PeV nu bkg ?

• In a model-indep. way, just assuming spectr. E-2 or E-2.2, the answer is yes - also suggesting a break @ few PeV

Murase, Ahlers, Lacki 2013, PRD, 88:121301

Some specific pp scenarios

• Radio Galaxies: CRs 10-100 EeV, escape into cluster IGM, where produce pp nus in the LSS ✔

• IGS (cluster accretion shocks): CRs @ 100PeV, then pp nu in IGM , & tdiff~tinj →sp. break ✔

• SBGs (starburst gals): may expect higher BISM, both SNe, HNe →CRs @ 100PeV, → pp nus ✔

• HNe (hypernovae) could be candidates (rate?)

Need: enough CR energy budget, pp efficiency

Murase, Ahlers, Lacki 2013, PRD, 88:121301

More speculatively:

• 3 ranges: LE (10-100GV); ME (PeV-EeV); UHE (10-100 EeV)

• UHE: ~1044 erg/Mpc3/yr

• ME: also ~1044 erg/Mpc3/yr

• LE: based on SBG radio and MW CR production: ~3x1045 erg/Mpc3/yr

• All three ~1044-1045, i.e. 1-dex flux in 10-dex energy

All CRs 109 -1021 eV from same type sources?

(Katz, Waxman, Thompson. Loeb, arX:1311.0287)

IF they are all the same sources, they need to be transients(to explain current UHECR flux lower than time average value)

Outlook for GRB• Observational prospects for space and ground-based multi-

GeV astrophysics of GRB are encouraging - there are enough bursts with enough photons statistics!

• Major theoretical issues: 1) baryonic vs. magnetic jets: how to distinguish them? 2) Prompt emission: hadronic or leptonic? 3) Are photospheres needed? Magn. or baryon. photospheres? 4) Are modified internal shocks competitive? Magnetized or hadronic internal shocks? 5) Reconnection or Fermi acceler’n? 6) MeV-GeV delay: astrophysical cause? 7) Pairs: important?

• TeV photon spectra, if detected, would provide new constraints on hadronic vs. leptonic models - but CTA may be needed, and much further theoretical work is needed to specify models

• Of course, GeV-TeV neutrino observations could clinch this issue - further address GRB-GZK UHECR contrib.

• Would constrain particle acceleration / shock parameters, emission region compactness (dimension, mag.field), etc.

Outlook for UHECR

• The sources of the UHECR are still unknown..!• They are almost certainly astrophysical sources (not TD)• GRB remain good candidates, as well as HNe, RQ AGN,

maybe MGRs.• Will increasingly constrain such possibilities with GeV

and TeV photon observations

• Will learn even more if & when astrophysical UHENUs are observed from any type of source

• Constraints from diffuse (and intrasource) γ-ray emission will also be very useful, and may remain for a long time the main constraint

• Composition and clustering will provide important clues

• A network of various complementary messenger facilities

• http://amon.gravity.psu.edu/amon_system.shtml

• Trigger: sub-threshold signal in 2 or more ≠ messengers

• See talk by Gordana Tešić (Friday)

Combining and exploiting the synergy between the four messengers:

IceCube Collaboration Meeting/Munich Fall 2013AMON

What AMON Will Do• Accept (sub-threshold) data from triggering observatories,

especially data that cannot be used standalone, e.g.:• single muon neutrinos from IceCube,

• two-interferometer coincidences from LIGO-Virgo,

• sub-threshold Swift BAT signals, etc.

• Streamline administrative overhead:

• Run temporal and spatial coincidence algorithms provided by users

• Issue “alerts” when coincidences are found to initiate closer investigation or OFU

• Help usher in the era of multimessenger astronomy!

64

→

status quofuture

IceCube Collaboration Meeting/Munich Fall 2013AMON

AMON Members

65

Triggering Observatories Follow-up Observatories

http://amon.gravity.psu.edu/participants.shtml

IceCube Collaboration Meeting/Munich Fall 2013AMON

Possible Sources

66

IceCube Collaboration Meeting/Munich Fall 2013AMON

Definition of Sub-Threshold• Threshold set low enough that false positive rate dominates total

event rate from observatory• acting alone, single facility cannot distinguish sub-threshold events from

background

• IceCube example:

67

data

IceCube Collaboration Meeting/Munich Fall 2013AMON

Do FoV’s Overlap Enough?

68

Average number of observatoriesviewing a source simultaneously

One-year simulation for IceCube, ANTARES, HAWC,Swift BAT, Auger, Fermi LAT and LIGO-Virgo

Number of triggering facilities observing asource (averages over time & sky position)

Yes.

IceCube Collaboration Meeting/Munich Fall 2013AMON

Current Software Status• Database constructed (clustered mySQL)

• populated (for now) with public data (IC40, Swift, Fermi)

• might switch to PostgreSQL

• Framework for coincidence algorithms in β

• Observatory⇄AMON network connections for being tested (http protocol)

69

IceCube Collaboration Meeting/Munich Fall 2013AMON

Analysis Chain

70

Event: trigger time, event coordinates, number of events, positional error or probability density function (PDF), observatory pointing and FPR

clustering analysis

Distribute candidate source position in real-time to follow-up facilities together with statistically valid FAR.

IceCube Collaboration Meeting/Munich Fall 2013AMON

Predicted AMON Gain for ν-γ

71

Up to 106 gain

Astropart. Phys. 45, 56-70, 2013

IceCube Collaboration Meeting/Munich Fall 2013AMON

Conclusions• AMON will provide a way to exploit data otherwise unusable

for astrophysical source searches• Example: it is much easier to get one ν from a source than two

• Starting now with archival data, AMON will provide a framework for multimessenger coincidence searches

• 2nd AMON workshop Oct 24-26, 2013 at Penn State• presentations from theorists and participating observatories

• data challenge using AMON framework and hidden signal

• Hope to graduate soon to real time operation...and discoveries!• Realtime use of future events would give outstanding discovery

potential

•72