Physical properties of the gamma-ray binary LS 5039 through low and high frequency radio observations

Benito Marcote

M. Ribó, J. M. Paredes, C. H. Ishwara-Chandra

HEPRO V October 7, 2015 Gamma-Ray Binaries

Gamma-ray binaries: binary systems which host a compact object orbiting a high mass that have the non-thermal maximum of the Spectral Energy Distribution in γ-rays (Paredes et al. 2013, Dubus 2013).

System Main star P/ days 0.10 LS 5039 O6.5 V 3.9 1FGL J1018.6–5856 O6 V 16.6 0.05 LS I +61 303 B0 Ve 26.5 0.00 HESS J0632+057 B0 Vpe 315.0

PSR B1259–63 O9.5 Ve 1236.7 AU 0.05 − Cygnus X-3 WR 0.2 0.10 Cygnus X-1 ?? O9.7 Ve 5.6 − MWC 656 ?? Be 60.4 0.15 −

0.15 0.10 0.05 0.00 0.05 0.10 Red: known gamma-ray binaries − − − Green: X-ray binaries with gamma-ray AU emission

2 /23 Benito Marcote [email protected] HEPRO V - La Plata 2/23 The gamma-ray binary LS 5039

h m s αJ2000 = 18 26 15.06 ◦ ′ ′′ δJ2000 = −14 50 54.3

O6.5 V main-sequence star (23  3 M⊙) 38 star Compact object, NS or BH (1–5 M⊙)

36 COMPTEL ≈ P 3.9 d RXTE EGRET BATSE

e = 0.35  0.04 34 SSC

[erg/s]) star IC ε corona  L d = 2.5 0.5 kpc ε sync. corona IC log(

32 int. Bremsstr. X-rays: periodic VLA ext. Bremsstr.

GeV light-curve: periodic (anticorrelated) 30 disk TeV light-curve: periodic disk IC

Radio: persistent, small variability 28 −5 −3 −1 1 3 5 7 9 11 13 without orbital modulation log (photon energy [eV]) FIGURE 1. Computed SED for LS 5039 compared with observational data. The arrows represent upper limits. Aharonian et al. (2005), Casares et al. SED of LS 5039 (Paredes et al. 2005) (2005), Kishishita et al. (2009), Abdo et This means that with reasonable parameter values, the assumptions of our model are consistent with the general features of the spectrum of LS 5039 from radio to TeV al. (2009), Casares et al. (2012), Zabalzagamma-ray energies. et al. (2013), Collmar & Zhang (2014) ACKNOWLEDGMENTS 3 /23 Benito Marcote [email protected] HEPROV.B-R. V - La and Plata J.M.P. acknowledge partial support by DGI of the spanish Ministerio3/23 de Educación y Ciencia (MEC) under grant AYA2004-07171-C02-01, as well as additional support from the European Regional Development Fund (ERDF/FEDER). V.B-R. is sup- ported by the DGI of the MEC under the fellowship BES-2002-2699. G.E.R is supported by the Argentine Agencies CONICET and ANPCyT (PICT 03-13291). This research has benefited from a cooperation agreement supported by Fundación Antorchas.

REFERENCES

1. Paredes, J.M., Martí, J., Ribó, M., & Massi, M., 2000, Science, 288, 2340 2. Strong, A. W., et al. 2001, AIP Conference Proceedings, 587, 21 3. Collmar, W., 2003, Proc. 4th Agile Science Workshop, p.177, Frascati (Rome) on 11-13 June 2003 4. Casares, J., Ribó, M., Ribas, I., Paredes, J.M. et al., 2005, MNRAS, submitted. 5. Bosch-Ramon, V., Paredes, J.M., Ribó, M. et al., 2005a, ApJ, 628, 388 6. Bosch-Ramon, V., Romero, G.E., & Paredes, J.M. 2005b, A&A, submitted The gamma-ray binary LS 5039

15 1 − s 2 h m s − αJ2000 = 18 26 15.06 10 ◦ ′ ′′ erg cm 12 δJ2000 = −14 50 54.3 − 10 / 5 10 keV − 1

O6.5 V main-sequence star (23  3 M⊙) F 0 ⊙ Compact object, NS or BH (1–5 M ) 1 3 − s 2 − cm

5 2 ≈ −

P 3.9 d 10 /

e = 0.35  0.04 1 30 MeV − 10 d = 2.5  0.5 kpc F 0 1.5 1 − s 2

X-rays: periodic − cm

6 1.0 GeV light-curve: periodic (anticorrelated) − 10 /

TeV light-curve: periodic 1 GeV 0.5 − 1 . 0 Radio: persistent, small variability F without orbital modulation 0.0

1 4 − s 2 − cm 12

Aharonian et al. (2005), Casares et al. − 2 10 (2005), Kishishita et al. (2009), Abdo et / 1 TeV > al. (2009), Casares et al. (2012), Zabalza F 0 et al. (2013), Collmar & Zhang (2014) 0.0 0.5 1.0 1.5 2.0 φ (P = 3.90603 d, MJD0 = 51942.59) 3 /23 Benito Marcote [email protected] HEPRO V - La Plata 3/23 Previous radio observations of LS 5039

• Variability from Martí et al. (1998) J. Mart´ı et al.: The system LS 5039: a new massive radio emitting X-ray binary L73

1.7 LS 5039 V-Band 1998 May 12

1998 Apr 09 -14 49 00 1.6

1998 Mar 10 LETTER

30 1.5 C1 1998 Feb 11 50 00

1.4 30 ROSAT LS5039 1.3 51 00 log Flux Density (mJy) (J2000) 30 1.2 C2

52 00 1.1 30 −0.2 0.0 0.2 0.4 0.6 0.8 1.0 1.2 log Frequency (GHz) 53 00 18 26 20 15 10 05 Spectral indexFig. 2. Theα = radio−0 spectrum.46  of0. LS01, 5039 variability on different< epochs25 during% the RIGHT ASCENSION (J2000) 1998 observations with the VLA. Error bars not shown are smaller than Fig. 3. CCD4 image of LS 5039 obtained on 1998 June 7 with the 1.52 the symbol size. /23 ′′ Benito Marcote [email protected] HEPRO V - La Plata m OAN telescope4/23 in the V -band. The circle with 35 radius represents the 90% confidence position from the ROSAT PSPC pointed observa- Table 2. Optical photometry of LS 5039. tions quoted in M97. The two comparison used for differential photometry are labeled as C1 and C2. Date Julian Day VRI − (JD 2400000) ground extragalactic source with S20cm ≃ 30 mJy, or brighter, ′′ < −7 1998 Jun 2 50966.5 11.35 −− within 0. 2 of the LS 5039 optical position is as low as ∼ 10 . 1998 Jun 3 50967.6 11.33 −− To our knowledge, this discovery increases to four the number 1998 Jun 7 50971.6 11.35 10.64 9.90 of confirmed massive X-ray binaries with associated radio emis- ◦ 1998 Jun 8 50972.5 11.39 10.69 9.95 sion, the other three being SS 433, Cygnus X-1 and LSI+61 303.

4.2. Radio behavior and they were reduced using standard procedures based on the IRAF image processing system. In Fig. 3 we show a 30 s expo- The radio spectra of LS 5039 in Fig. 2 are very suggestive of non- sure of the LS 5039 field in the V -band. Differential photometry thermal radio emission. A typical one can be well represented by −0.46±0.01 was performed against two nearby comparison stars (C1 and C2 Sν = (52±1) mJy (ν/1 GHz) . Given the unresolved ′′ in Fig. 3). angular size (θ ≤0. 1), the brightness temperature Based on several observations of Landolt et al. (1992) stan- −2 −2 3 ν θ S dards, the adopted magnitudes of the comparison stars are T =1.76 × 10 K ν (1) b GHz #arcsec$ #mJy $ V = 12.53, R = 12.01, I = 11.55 and V = 10.10, R =9.36, ! " I =8.68 for C1 and C2, respectively. This absolute photom- can be estimated, for example at the 20 cm wavelength, as 6 etry is accurate to ±0.03 mag. Relative variations in the final Tb ≥ 4 × 10 K. Such a high value, together with the signifi- photometric results of Table 2 can be nevertheless traced at the cantly negative spectral index, safely rules out a thermal emis- ±0.01 mag level. sion mechanism. Non-thermal synchrotron radiation remains therefore as the most plausible interpretation for the LS 5039 radio emission. We can also obtain a lower limit to the source 4. Discussion angular size by preventing catastrophic inverse Compton losses 12 4.1. Astrometric identification to occur (Tb ≤ 10 K). This condition yields θ ≥ 0.2 mas, implying that further progress in studying the LS 5039 struc- The radio position derived from the VLA-A configuration maps ture may be feasible by the use of VLBI techniques. We ignore h m s ± s in Fig. 1 is found to be: αJ2000 = 18 26 15. 056 0. 001 at present if the radio emission is originated in collimated mil- − ◦ ′ ′′ ± ′′ and δJ2000 = 14 50 54. 24 0. 01. These coordinates agree liarcsec radio jets or, alternatively, some other mechanism is at within 0.′′1 with those of LS 5039 in the optical USNO-A1.0 ′′ work. catalogue, whose typical astrometric error is about 0. 25. There- Assuming equipartition arguments for synchrotron radio fore, we conclude that the identification of the LS 5039 radio sources (Pacholczyk 1970), the total energy content and mag- counterpart on the basis of astrometrical coincidence at the sub- netic field in LS 5039 can be expressed as: arcsec level is virtually certain. From source count analysis (e.g. 4/7 3/7 9/7 4/7 Ledden et al. 1980), the a priori probability of having a back- Etotal = c13(1 + k) φ R Lrad (2) Understanding the emission of Gamma-Ray Binaries Young non-accreting pulsar scenario V. Zabalza et al.: GeV-TeV emission model for LS 5039 Strong shock between both winds: The pulsar wind termination region close to the wind stand- off location has been generally considered as the preferred loca- • tion for acceleration of electrons up to ultrarelativistic energies, Relativistic pair plasma wind from which would give rise to the observed non-thermal broadband emission through synchrotron and IC processes (e.g., Dubus, the pulsar 2006). However, the intense stellar photon field at this location, Coriolis and corresponding high opacities owing to photon-photon pair • Stellar wind from the massive turnover production, precludes VHE gamma-rays above 40 GeV from reaching the observer for some orbital phases with⇠ clear VHE companion star Orbital detections (Khangulyan et al., 2008a). On the other hand, an motion ind emitter with constant particle injection located at the wind ar w puls standoff is compatible with the phenomenology exhibited by Originally proposed by Maraschi & Treves cked Pulsar Sho the source at GeV energies. The modulation is mainly driven by Wind (1981), re-proposed by Dubus (2006) standoff the change in the IC interaction angle, with enhanced (reduced) Shocked stellar wind emission at superior (inferior) conjunction. For the allowed angles (20 –60 , Casares et al., 2005), the modulation is too strong to reproduce the Fermi lightcurve, but Star this effect may be mitigated by Doppler boosting. As a proxy for • High energy emission produced by the complex hydrodynamical properties of the flow (Bogovalov et al., 2008), we have considered an effective bulk flow velocity synchrotron and inverse Compton. of 0.15c in the direction of the axis of symmetry of the CD apex, which changes along the (see CD aberration in Sect. 2) • Radio flux dominated by the but is always close to the radial direction. The bulk flow veloc- Model for LS 5039 (Zabalza et al. 2012) ity of 0.15c used here applies exclusively to the Doppler effect averaged over the CD cone and with respect to the observer synchrotron emission (Bosch-Ramon Fig. 1. Sketch of the proposed scenario at the periastron of a line-of-sight. Therefore, it is still possible, and, given the sim- close-orbit system similar to LS 5039. The region of the CD ulations of Bogovalov et al. (2008), probable, that the intrinsic 2009) where significant turbulence, and therefore wind mixing, are bulk flow velocity of the shocked pulsar wind is higher. expected to take place is indicated with a wavy line. The red The pulsar wind termination shock at the Coriolis turnover regions indicate the two proposed emitter locations. location is a good candidate for the VHE emitter. The reduced 5 stellar photon field density implies pair production absorption /23 Benito Marcote [email protected] HEPRO V - La Plata 5/23 characteristics consistent with the observed spectra at VHE. For simulations of Bosch-Ramon et al. (2012), and will therefore be scenarios in which the stellar wind ram pressure is dominant, used throughout this work. the pulsar wind solid angle subtended by the Coriolis turnover The presence and properties of this quasi-perpendicular shock is much lower than that of the apex region. This is con- strong shock at the Coriolis turnover location sets apart the sistent with the reduced emission power of the TeV component characteristics of the interaction of stellar and pulsar wind with with respect to the GeV component. respect to that of binary stellar systems. The non-relativistic The magnetic field of the shocked pulsar wind is gener- stellar wind in binary stellar systems can be smoothly deflected ally well known in plerions from observations (e.g., Kennel & by sound waves, as found in the simulations of Lamberts et al. Coroniti, 1984), but the so-called sigma problem precludes one (2012). However, sound wave deflection would not affect an from deriving it theoretically. The situation in gamma-ray bi- already trans-sonic relativistic pulsar wind, leading to the for- naries is even more complex, as there might be a reduction of mation of a quasi-perpendicular strong shock because of struc- the post shock magnetic field strength owing to the reaccel- ture bending. We note that this approach may not be valid for eration of the flow (Bogovalov et al., 2012). Since knowledge 1.2 10 2, as numerical simulations indicate that a recon- of the post-shock magnetic field could only be obtained via a ⌘w Æ finement⇥ shock, with different characteristics to those explored magnetohydrodynamical simulation of the system, we have above, will develop in the pulsar wind for those cases where chosen to parametrize the magnetic field energy density as a the stellar wind vastly overpowers the pulsar wind (Bogovalov fraction ⇠ of the pre-shock equipartition magnetic field. We et al., 2008). define the latter at the balance of the magnetic field energy density and the kinetic energy density of the pulsar wind at the shock. Therefore, for a constant value of ⇠, the post-shock magnetic field strength will behave along the orbit as B 1/r , 3. Location and properties of the non-thermal emitters p where rp is the distance of the shock to the pulsar. / A sketch of the system, as described in the previous section, can be found in Fig. 1. The influence of the orbital motion on the 3.1. Maximum electron energy shape of the pulsar wind zone gives rise to two distinct regions where most of the energy of the pulsar wind is deposited in A distinct feature of the GeV and TeV spectra of LS 5039 are the quasi-parallel shocks: the wind standoff and Coriolis turnover cutoffs present at a few GeV and above 10 TeV, respectively. The locations. The intermediate regions of the pulsar wind shock, former is apparently constant along the orbit, whereas owing i.e., those not marked in red in Fig. 1, are not expected to con- to the softer TeV spectrum during superior conjunction the lat- tribute significantly to the non-thermal emission of the system ter has only been detected clearly during inferior conjunction. given the high obliquity of the impact of the pulsar wind and These spectral cutoffs can be directly related to the maximum correspondingly low energy transfer into non-thermal parti- energy of the electrons in the respective particle populations. cles. The maximum energy that can be reached in any acceleration

3 Previous radio observations of LS 5039

• VLBI observations: Moldón et al. (2012)

Observations along one . Dominant core emission (≲ 1 mas, or 3 AU). Extended emission orbitally modulated at mas scales (<10% of the total flux density emission).

6 /23 Benito Marcote [email protected] HEPRO V - La Plata 6/23 Previous radio observations of LS 5039

• At low frequencies, only a few observations have been published with contradictory results. L2 S. Bhattacharyya et al.

synchrotron100 self-absorption (SSA) was responsible for the optically thick radio spectrum, then the magnetic field wasPandey found et to al. be(2007) in A 2 Pandey et al. (2007) B the range 10− 1G. The size of the emitting region would be −13 Mart´ı et al. (1998) (4.5–10)50 10 cm and the location of the source would be at ∼ × 13 !(4.5–10) 10 cm. Hence it is important to understand the actual process responsible× for the absorption of radio emission. Here we present the results of radio observations of LS 5039 taken with the Giant Metrewave Radio Telescope (GMRT) in the 200– 1280 MHz frequency range. We study the low-frequency spectrum during the periastron and apastron passage of the compact star and attemptFlux density (mJy) to understand its possible consequences. The observation

and10 the data analysis are discussed in the next section. The results are discussed in Section 3, and finally we conclude the Letter in Section 4.

5 2OBSERVATIONSANDDATAREDUCTION100 1000 10000 Frequency (MHz) The GMRT consists of 30 steerable antennas of 45-m diameter in an approximate ‘Y’ shape similar to the VLA but with each antenna Figure 1.BhattacharyyaRadio spectrum of LS 5039 et as observed al. (2012) with the GMRT. The Adapted from Pandey et al. (2007) archival data are also plotted. in a fixed position. 14 antennas are randomly placed within a central 1 1km2 (the ‘central square’) and the remaining antennas form Godambe et al. (2008) baseline interferometry (VLBI) resolved the source and we have the× irregular Y shape (six on each arm) over a total extent of about used the flux values reported for the core. The overall flux from 25 km. We refer the reader for details about the GMRT array to the core and the wings is of the same order as reported by Mart´ı http://gmrt.ncra.tifr.res.in and Swarup et al. (1991). We observed et al. (1998). It is evident that the spectra remain optically thick7 LS 5039 at 1280 MHz on 2006 February 23 and simultaneously and the flux values at frequencies 605 and 1280 MHz are almost/23 Benitoobserved Marcote at 234 and [email protected] 614 MHz on 2006 March 13. HEPRO V - La Plata 7/23 equal during the periastron (phase range: 0.65–0.72) and apastron The observations were made in standard fashion, with each source (phase range: 0.99–0.06) passages. Because of the poor quality of observation (30 min) interspersed with observations of the phase the data at 234 MHz, we could not produce a source image for both calibrator (4 min). The primary flux density calibrator was either the observing spells. Compared to the flux values of our previous 3C 48 or 3C 286, with all flux densities being on the scale of Baars observations, the present flux values fall within the error bars of the et al. (1977). Either of these flux calibrators was observed for 20 min previous observations. It is to be noted that the exposure time for our at the beginning and end of each observing session. previous observations was less than that of the present observations. The data recorded with the GMRT were converted to FITS for- The optically thick spectrum at radio frequencies can be produced mat and analysed with the Astronomical Image Processing System due to the absorption of radiation either by free–free absorption or by (AIPS) using standard procedures. During the data reduction, we SSA in the emitting region of the source. The free–free absorption had to discard a fraction of the data affected by radio frequency of radio waves may occur in LS 5039 due to the presence of stellar interference and system malfunctions. After editing, the data were wind from the primary star. The relativistic jet where the radio calibrated and collapsed into fewer channels. A self-calibration on emission takes place by the synchrotron process moves through an the data was used to correct for the phase-related errors and improve environment of such a stellar wind. For the time being, let us assume the image quality. The observation dates, the observed frequencies that the wind from the companion star is spherically symmetric. If and the corresponding spectral phases are given in Table 1. so, the free–free absorption opacity is given by (Rybicki & Lightman 1986)

3RESULTSANDDISCUSSION ff (3/2) 2 2 2 α 0.018T − Z n ν− g¯ , (1) ν = ff The spectra obtained for the two orbital phase ranges corresponding where T is the temperature of the stellar wind, n is the number to the periastron and apastron passage are shown in Fig. 1 along density of the stellar wind and ν is the frequency of radiation. with the data points from our previous observations (Paper I) and For a binary orbit inclined at an angle i with respect to the sky archival data points from Mart´ı, Paredes & Ribo´ (1998) and Ribo´ plane, the jet will make an angle i with the direction of observation. et al. (2008). The flux points obtained from Ribo´ et al. (2008) do It is assumed that the jet is perpendicular to the orbital plane. If the not match with the points obtained by Mart´ı et al. (1998). This semimajor axis of the orbit is a and the eccentricity is e, then the is because the observation by Ribo´ et al. (2008) with very long equation of the orbit is given by Table 1. Log of GMRT observa- 2 2 2 a (1 e ) tions. ρ − , (2) = (1 e2 cos2 θ) − MJD Frequency Phase where θ is the phase angle and θ 2πφ,0 φ 1 is the orbital (MHz) range phase. From Fig. 2, using simple geometry,= one≤ can≤ write

2 2 2 2 54573 234 0.63–0.70 O′C ρ a e 2ρae cos θ. (3) 605.2 0.63–0.70 = + − 54670 1280 0.65–0.72 If the radio-emitting region is located at a height z from the base of 54684 1280 0.98–0.05 the jet, then 54688 605.2 0.99–0.06 2 2 2 2 2 2 2 s O′C z ρ a e 2ρae cos θ z . (4) = + = + − +

C ⃝ 2011 The Authors, MNRAS 421, L1–L5 C Monthly Notices of the Royal Astronomical Society ⃝ 2011 RAS Radio observations Work published in Marcote et al. (2015)

• 1.4 – 15 GHz (VLA) Monitoring in 1998, 2002

• 154, 235 & 610 MHz (GMRT) Archival observations 2004–08

• 154 MHz – 5 GHz (GMRT & WSRT) Two quasi-simultaneous observations in 2013.

8 /23 Benito Marcote [email protected] HEPRO V - La Plata 8/23 Summary of the observations analyzed in this work

9 /23 Benito Marcote [email protected] HEPRO V - La Plata 9/23 Summary of the observations analyzed in this work

9 /23 Benito Marcote [email protected] HEPRO V - La Plata 9/23 High-frequency variability along the orbit VLA monitoring in 2002

• Persistent flux density emission. • Variability < 25 % • Variability on timescales as • No visible orbital modulation short as one day. Flux density values at all frequencies compatible with the ones reported at other epochs (e.g. Martí et al. 1998).

10 /23 Benito Marcote [email protected] HEPRO V - La Plata 10/23 High-frequency variability along the orbit VLA monitoring in 2002

• Dashed lines represent the power-laws determined from 5.0 and 8.5 GHz

• We observe a ≈ power-law at these GHz frequencies most of the times

• Slight curvature below 2 GHz

• Average spectral index: α = −0.57  0.12

11 /23 Benito Marcote [email protected] HEPRO V - La Plata 11/23 Low-frequency variability along the orbit Archival GMRT observations 2004–2008

12 /23 Benito Marcote [email protected] HEPRO V - La Plata 12/23 Low-frequency variability along the orbit Archival GMRT observations 2004–2008

• Observations spread over 4 yr

• Persistent emission

• Variability at > 6σ

• Average spectral index α = +0.5  0.8

• Hints of orbital modulation?

12 /23 Benito Marcote [email protected] HEPRO V - La Plata 12/23 Non-simultaneous spectrum of LS 5039 Combining data from 1998 to 2013

50 • Small variability along the (< 25 % ∀ ν) 20 mJy /

• ⌫ “Similar” profile in average S 10 VLA 1998 • Turnover at ∼ 0.5 GHz VLA 2002 5 GMRT 2004–2008 GMRT-WSRT 2013 • Source undetected at 150 MHz 0.2 0.5 1 2 5 10 20 50 None • Only two data at 2.3 GHz (no statistics) 20 mJy / ⌫

• The mean square errors have S 10 been used in the average data 5

0.2 0.5 1 2 5 10 ⌫/GHz 13 /23 Benito Marcote [email protected] HEPRO V - La Plata 13/23 Modeling the LS 5039 spectrum A first approximation (toy model)

From VLBI observations (Moldón et al. 2012) we known that most of the radio emission comes from a compact core ≲ 1 mas (∼ 3 AU) to be compared with the 0.19 AU of the semi major axis.

We have built a very simple model to understand the spectrum: • Compact core ; one-zone model • No orbital modulation ; symmetric emitting region (spheric) • For simplicity ; isotropic and homogeneous • We consider the presence of a synchrotron emitting plasma • Turnover produced either by SSA, FFA or Razin effect (and combinations of them)

14 /23 Benito Marcote [email protected] HEPRO V - La Plata 14/23 Modeling the LS 5039 spectrum A first approximation (toy model) We have built a very simple model to understand the spectrum: • Synchrotron emission, with a particle injection:

n(E)dE = KE−pdE

• Synchrotron self-absorption (SSA):

(p+2)/2 −(p+4)/2 κSSA ∝ KB ν

• Free-free absorption (FFA):

∝ 2 −3/2 −2 κFFA neTw ν

• Razin effect:

−νR/ν −1 Sν ; Sνe , νR ≡ 20neB

15 /23 Benito Marcote [email protected] HEPRO V - La Plata 15/23 Non-simultaneous spectrum of LS 5039 Combining data from 1998 to 2013

• The average spectrum can be fitted by typical models: ◦ SSA ◦ Synchrotron + FFA ◦ SSA + Razin ◦ FFA + Razin ◦ SSA + FFA • SSA+Razin effect is the best fit to the data • But the other fits are not statistically rejected • Small differences between all of them

16 /23 Benito Marcote [email protected] HEPRO V - La Plata 16/23 Quasi-simultaneous spectrum of LS 5039 GMRT & WSRT campaign in 2013 July 19 and 21

17 /23 Benito Marcote [email protected] HEPRO V - La Plata 17/23 Quasi-simultaneous spectrum of LS 5039 GMRT & WSRT campaign in 2013 July 19 and 21

Two 0.15–5 GHz spectra at orbital phases ϕ ≈ 0.9 and 0.4.

• Although similar to the average spectrum, we observe subtle differences between the two epochs. • The turnover at ∼ 0.5 GHz is persistent. • Stronger emission on 2013 July 19. • 2013 July 19: pure SSA spec. GMRT data: 235 and 610 MHz WSRT data: 2.3 and 5.0 GHz • 2013 July 21: SSA+Razin spec. 154 MHz GMRT data taken every other day • FFA provides poor fits (on 2013 July 18, 20 and 22)

17 /23 Benito Marcote [email protected] HEPRO V - La Plata 17/23 Modeling the LS 5039 spectrum

The three spectra show a similar shape but with subtle differences: • Avg. spectrum: SSA+Razin • July 19 spectrum: SSA • July 21 spectrum: SSA+Razin

Fit p ΩB−1/2 KℓB(p+2)/2 ν [ ][ ] [ R ] 10−16 G−1/2 10 3 cm G(p+2)/2 10 8 Hz

Avg. spectrum 2.16  0.04 500  800 3  5 4.1  0.2 July 19 1.867  0.014 3.9  0.3 (2.1  0.9) × 106 – July 21 2.24  0.08 200  600 0.4  1.7 4.1  0.7

• We can also compare these results with the free-free opacity inferred from the stellar wind (the region must be optically thin to FFA)

18 /23 Benito Marcote [email protected] HEPRO V - La Plata 18/23 Modeling the LS 5039 spectrum Building a coherent picture from the fits and the free-free opacity: • Avg. spectrum: SSA+Razin • July 19 spectrum: SSA • July 21 spectrum: SSA+Razin • Coherent picture with: ℓ ∼ 0.85 mas (∼ 2.5 AU) B ∼ 20 mG 5 −3 ne ∼ 4 × 10 cm −8 −1 M˙ ∼ 5 × 10 M⊙ yr where: ℓ: linear size of the emitting region, B: module of the magnetic field, ne: electron density of the non- relativistic plasma, M˙ : mass-loss rate of the companion star. 19 /23 Benito Marcote [email protected] HEPRO V - La Plata 19/23 Modeling the LS 5039 spectrum Building a coherent picture from the fits and the free-free opacity: • Avg. spectrum: SSA+Razin • July 19 spectrum: SSA • July 21 spectrum: SSA+Razin • Coherent picture with: ℓ ∼ 0.85 mas (∼ 2.5 AU) B ∼ 20 mG 5 −3 ne ∼ 4 × 10 cm −8 −1 M˙ ∼ 5 × 10 M⊙ yr where: ℓ: linear size of the emitting region, B: module of the magnetic field, ne: electron density of the non- relativistic plasma, M˙ : mass-loss rate of the companion star. 19 /23 Benito Marcote [email protected] HEPRO V - La Plata 19/23 Modeling the LS 5039 spectrum Building a coherent picture from the fits and the free-free opacity: • Avg. spectrum: SSA+Razin • July 19 spectrum: SSA • July 21 spectrum: SSA+Razin • Coherent picture with: ℓ ∼ 0.85 mas (∼ 2.5 AU) B ∼ 20 mG 5 −3 ne ∼ 4 × 10 cm −8 −1 M˙ ∼ 5 × 10 M⊙ yr where: ℓ: linear size of the emitting region, B: module of the magnetic field, ne: electron density of the non- relativistic plasma, M˙ : mass-loss rate of the companion star. 19 /23 Benito Marcote [email protected] HEPRO V - La Plata 19/23 Modeling the LS 5039 spectrum Building a coherent picture from the fits and the free-free opacity: • Avg. spectrum: SSA+Razin • July 19 spectrum: SSA • July 21 spectrum: SSA+Razin • Coherent picture with: ℓ ∼ 0.85 mas (∼ 2.5 AU) B ∼ 20 mG 5 −3 ne ∼ 4 × 10 cm −8 −1 M˙ ∼ 5 × 10 M⊙ yr where: ℓ: linear size of the emitting region, B: module of the magnetic field, ne: electron density of the non- relativistic plasma, M˙ : mass-loss rate of the companion star. 19 /23 Benito Marcote [email protected] HEPRO V - La Plata 19/23 Modeling the LS 5039 spectrum

• A significant mixing of the non-relativistic wind inside the synchrotron radio emitting relativistic plasma, even close to ∼ 100%, is observed.

• The derived mass-loss rate (model dependent, in agreement with the last results, Casares, in prep.) implies that the wind is clumpy.

• The presence of Razin effect, widely observed in Colliding Wind Binaries, could give further support to the scenario of the young non-accreting pulsar.

20 /23 Benito Marcote [email protected] HEPRO V - La Plata 20/23 Conclusions Marcote et al. (2015), MNRAS, 451, 59 Marcote (2015), PhD Thesis, Universitat de Barcelona

• We report day to day variability, trends on week timescales, and the absence of orbital variability. Variability < 25 % is observed at all frequencies (0.23–15 GHz) even on timescales. • Persistent turnover at around ∼ 0.5 GHz. • No detection up to now at 150 MHz • The considered simple model can explains the observed spectra, indicating that the turnover is dominated by SSA • A contribution of Razin effect is also observed at some epochs. • Even with this simple model we have obtained a coherent picture that explains the observed spectra. • Future multifrequency campaigns with more accurate results (specially at low frequencies) are required to compare with more detailed models.

21 /23 Benito Marcote [email protected] HEPRO V - La Plata 21/23 Orbital and superorbital variability of LS I +61 303 at low radio frequencies with GMRT and LOFAR B. Marcote1, M. Ribó, J. M. Paredes, C.H. Ishwara-Chandra, J. D. Swinbank, J. W. Broderick, S. Markoff, R. Fender, R. A. M. J. Wijers, G. G. Pooley, A. J. Stewart, M. E. Bell, R. P. Breton, D. Carbone, S. Corbel, J. Eislöffel, H. Falcke, J.-M. Grießmeier, M. Kuniyoshi, M. Pietka, A. Rowlinson, M. Serylak, A. J. van der Horst, J. van Leeuwen, M. W. Wise, P. Zarka 1Departament d’Astronomia i Meteorologia, Institut de Ciències del Cosmos (ICCUB), Universitat de Barcelona (IEEC-UB)

LS I +61 303 is a gamma-ray binary that exhibits an outburst at GHz frequen- cies each orbital cycle of 26.5 d and a superorbital modulation with a period of 4.6 yr. We have performed a detailed study of the low-frequency radio emission of LS I +61 303 by analysing all the archival GMRT data at 150, 235 and 610 MHz, and conducting regular LOFAR observations within the Radio Sky Monitor (RSM) at 150 MHz. We have detected the source for the first time at 150 MHz, which is also the first detection of a gamma-ray binary at such a low frequency. We have obtained the light-curves of the source at 150, 235 and 610 MHz, all of them showing orbital modulation. The light-curves at 235 and 610 MHz also show the existence of superorbital variability. A comparison with contemporaneous 15-GHz data shows remarkable differences with these light-curves. At 15 GHz we see clear outbursts, whereas at low frequencies we see variability with wide maxima. The light-curve at 235 MHz seems to be anticorrelated with the one at 610 MHz, implying a shift of ∼0.5 orbital phases in the maxima. We model the shifts between the maxima at different frequencies as due to changes in the physical parameters of the emitting region assuming either free-free absorption or synchrotron self-absorption, obtaining expansion velocities for this region close to the stellar wind velocity with both mechanisms. The inferred values would give further support to the young non-accreting pulsar wind scenario. All these results are published in Marcote et al. (2015, submitted).

300

250

200

150

Thank You…! (J2000) δ 100

050 GMRT LOFAR

+61◦000

…and see also the poster about 42m 41m 40m 2h39m 42m 41m 40m 2h39m α (J2000) α (J2000) Detection of LS I +61 303 in the 2008 Feb 18 154-MHz GMRT observation (left) and in the 2013 July 14 149-MHz LOFAR observation after removing baselines similar studies conducted for the below 0.2 kλ (right).

Field of LS I +61 303 seen by POSS II (Reid et al. 1991) at optical red F band gamma-ray binary LS I +61 303 (top), and by LOFAR at 149 MHz on 2013 July 14 (bottom). (no. 20)!

Flux density values of LS I +61 303 as a function of the MJD (left) and folded with the orbital period (right) from the analyzed GMRT and RT data. The black data (left bottom) shows the spectral index in the range of 235–610 MHz. We observe an orbital modulated light-curve. The 610-MHz data shows a quasi-sinusoidal modulation with a slower decay with respect to the 15-GHz data. The 235-MHz light-curve is almost anticorrelated with the one observed at 610 MHz.

160 LOFAR 150 MHz Folded light-curve of LS I +61 303 GMRT 150 MHz with the orbital period from the 150- 140 OVRO 15 GHz MHz data. We observe an or- Expected shifts in orbital phase (∆φ) 120 bitally modulated emission, with the for the maxima in the flux density 100 outburst slightly delayed with re- emission between a frequency ν and 15 GHz for different velocities mJy spect to the contemporaneous 15- /

ν 80 S GHz data. We note that these 150- of expansion of the radio emitting Inferred expansion velocity, assuming 60 MHz observations were taken at dif- region considering synchrotron FFA, as a function of the mass-loss rate. self-absorption with different de- 40 ferent superorbital phase (φso ∼ 0:8) The vertical dashed line denotes the ex- compared to the ones at 235 and pendences of the magnetic field with pected mass-loss rate. We compare the 20 610 MHz (0.2). Therefore we can not the size of the emitting region (νSSA, obtained velocities with the mean stellar 0 top figures) or free-free absorption 0.0 0.5 1.0 1.5 2.0 directly compare the behavior of the wind velocity (solid line) and its lower ex- φorb (Porb = 26.4960 d, JD0 = 2 443 366.775) source from these two data sets. (νFFA, top right column). pected value (dotted line).

22 LATEX TikZposter /23 Benito Marcote [email protected] HEPRO V - La Plata 22/23 Back-up slides

22 /23 Benito Marcote [email protected] HEPRO V - La Plata 22/23 Accurate GMRT data reduction (Marcote et al. 2015, Appendix) Analyzing the GMRT data we found the reasons of the differences between Pandey et al. (2007) and Godambe et al. (2008): • At low frequencies, the contribution of the Galactic diffuse emission is quite high within the Galactic Plane. • It must be removed to recover the right flux densities in the final image. Otherwise the flux densities will be underestimate. • Usually done automatically in most telescopes. Not in the GMRT. • Godambe et al. (2008) did not take into account this correction. • Pandey et al. (2007) used the Haslam approximation (see Marcote et al. 2015), extrapolating the emission seen at 408 MHz. • We have conducted dedicated non correlated observations with the GMRT to directly measure the Galactic contribution in the field of LS 5039 to properly substract it. • We have seen significant differences in these two methods forthe field of LS 5039 (compatible with the comparison in Sirothia 2009). 23 /23 Benito Marcote [email protected] HEPRO V - La Plata 23/23 Razin effect The synchrotron emission propagates through a plasma, which presents a refractive index n. Always that n < 1 the beaming effect is partially suppressed. Although at high frequencies the effect is negligible, at low frequencies (with ν ≪ νp, where νp is the plasma frequency) it suppresses the beaming effect since ( ) ν 2 n2 = 1 − p ν and the beaming effect goes as √ −1 2 2 θb ≈ γ = 1 − n β

• Presence of a thermal plasma surrounding the emitting region. • Attenuation of the synchrotron radiation at low frequencies. • Widely reported in Colliding Wind Binaries, solar wind,… • A good approximation is an exponential attenuation at low frequencies (Dougherty et al. 2003):

−νR/ν −1 Sν ∝ Sνe , νR ≡ 20neB 23 /23 Benito Marcote [email protected] HEPRO V - La Plata 23/23