Äîêóìåíò âçÿò èç êýøà ïîèñêîâîé ìàøèíû. Àäðåñ îðèãèíàëüíîãî äîêóìåíòà : http://star.arm.ac.uk/preprints/2010/556.pdf Äàòà èçìåíåíèÿ: Wed Jan 13 13:41:38 2010 Äàòà èíäåêñèðîâàíèÿ: Tue Oct 2 02:53:24 2012 Êîäèðîâêà: Ïîèñêîâûå ñëîâà: gro j1655-40

Mon. Not. R. Astron. Soc. 000, ??­?? (??)

Printed 13 January 2010

A (MN L TEX style file v2.2)

Disentangling jet and disc emission from the 2005 outburst of XTE J1118+480
C. Brocksopp1 , P.G. Jonker2,3, D. Maitra4,5, H.A. Krimm6,7, G.G. Pooley8, G. Ramsay9, C . Z ur i t a 1 0 1

arXiv:1001.1965v1 [astro-ph.HE] 12 Jan 2010

Mul lard Space Science Laboratory, University Col lege London, Holmbury St. Mary, Dorking, Surrey RH5 6NT SRON, Netherlands Institute for Space Research, 3584 CA Utrecht, The Netherlands 3 Harvard­Smithsonian Center for Astrophysics, 60 Garden Street, Cambridge, MA 02138, U.S.A. 4 Astronomical Institute "Anton Pannekoek" and Center for high-Energy Astrophysics, University of Amsterdam, Kruislaan 403, Amsterdam 1098 SJ, The Netherlands 5 Department of Astronomy, University of Michigan, Ann Arbor, Michigan 48109, USA 6 NASA/Goddard Space Flight Centre, Greenbelt, MD 20771, USA 7 Universities Space Research Association, Columbia, MD 20144, USA 8 Mul lard Radio Astronomy Observatory, Cavendish Laboratory, Madingley Road, Cambridge CB3 0HE 9 Armagh Observatory, Col lege Hil l, Armagh BT61 9DG 10 Instituto de Astrofisica de Canaris, 38200 LaLaguna, Tenerife, Spain
2

Accepted ??. Received ??

ABSTRACT

The black hole X-ray transient, XTE J1118+480, has now twice been observed in outburst -- 2000 and 2005 -- and on both occasions remained in the low/hard X-ray spectral state. Here we present radio, infrared, optical, soft X-ray and hard X-ray observations of the more recent outburst. We find that the lightcurves have very different morphologies compared with the 2000 event and the optical decay is delayed relative to the X-ray/radio. We attribute this lesser degree of correlation to contributions of emission from multiple components, in particular the jet and accretion disc. Whereas the jet seemed to dominate the broadband spectrum in 2000, in 2005 the accretion disc seems to be more prominent and we use an analysis of the lightcurves and spectra to distinguish between the jet and disc emission. There also appears to be an optically thin component to the radio emission in the 2005 data, possibly associated with multiple ejection events and decaying as the outburst proceeds. These results add to the discussion that the term "low/hard state" covers a wider range of properties than previously thought, if it is to account for XTE J1118+480 during these two outbursts. Key words: stars: individual: XTE J1118+480 -- accretion, accretion discs -- Xrays: binaries

1

INTRODUCTION

XTE J1118+480 is a black hole X-ray transient, notable for b eing one of a small subset of transient systems observed in the low/hard X-ray sp ectral state throughout at least one outburst (e.g. Brocksopp et al. 2004). The low/hard state is defined in terms of the X-ray sp ectrum; traditionally the Xray emission at 2­20 keV could b e well-describ ed by a p owerlaw, with the photon index, , in the range 1.5 < < 2.1 and with no or little need for a disc comp onent (McClin-

tock & Remillard 2006). This is consistent with the X-ray emission in this energy range b eing produced via Compton up-scattering of soft photons by a hot electron plasma (e.g. corona, jet or advection-dominated accretion flow) and, indeed, a partially self-absorb ed radio jet has now long b een associated with the low/hard state (e.g. Fender 2006 and references therein). Controversially, XTE J1118+480 then b ecame the first source for which a compact jet model could b e fit to the whole sp ectrum, from radio to hard X-rays, and required a disc contribution only in the optical and ultraviolet regions (Markoff, Falcke, Fender 2001; see also, however, alternative

email: cb4@mssl.ucl.ac.uk c ?? RAS


2

Brocksopp et al.
Table 1. Hard and soft X-ray count rates from Swift (top) and RXTE (bottom). The energy ranges for the BAT, PCA and HEXTE are 14­195, 3­20 and 20­200 keV respectively. MJD 53380 53382 53384 53388 53390 53392 53394 53396 53398 53404 MJD 53383. 53384. 53385. 53385. 53385. 53386. 53386. 53387. 53387. 53388. 53388. 53389. 53389. 53390. 53390. 53391. 53392. 53392. 53392. 53393. 53394. 53394. 53394. 53395. 53395. 53396. 53398. 53398. 53398. 53399. 53400. 53400. 53400. 53403. 3 2 0 7 8 4 8 2 8 6 8 1 6 1 5 1 0 1 7 7 2 2 7 1 6 1 0 1 4 3 4 5 9 4 BAT (cts/s/cm2 ) 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0024 0073 0102 0067 0058 0057 0033 0020 0020 0012 ± ± ± ± ± ± ± ± ± ± 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0006 0003 0003 0007 0003 0006 0007 0007 0004 0007

explanations in Zdziarski et al. 2003, Zdziarski & Gierlinski 2004 and revised jet models in e.g. Markoff, Nowak & Wilms 2005, Maitra et al. 2009). The implications of a dominant X-ray synchrotron and/or inverse Compton emission comp onent for the energy budget of the jet are far-reaching and, if confirmed, would necessitate significant changes to current models for accretion discs and their outbursts. Similar jet models have b een used to describ e low/hard state sp ectra of GX 339-4 and Cyg X-1 (Markoff, Nowak, Wilms 2005). However, GX 339-4, Cyg X-1 and the 2000 outburst of XTE J1118+480 have had unusually long p eriods when a stable jet is present; the presence of jets is more commonly restricted to brief phases of a transient outburst (Fender, Belloni, Gallo 2004). In such cases the canonical self-absorb ed radio sp ectrum of the low/hard state is likely to b e "contaminated" with optically thin radio emission (e.g. Jonker et al. 2009). While we have a reasonably good understanding of the sequence of progression from one sp ectral state to another, evolution of the jet over the course of a low/hard state outburst has not yet b een studied in detail. We do not yet know whether a jet, emitting from radio to X-ray wavelengths, is a typical feature of transient outbursts or a rare phenomenon, occuring only under extreme conditions in some sp ecific sources or events. The relative imp ortance of jet, disc and corona is one of the key questions currently facing the field of X-ray binaries. Markoff et al. (2005) show that high signal-to-noise sp ectra of Cyg X-1 and GX 339-4 can b e equal ly well fit by jet and Compton models. Hynes et al. (2006) emphasised the difficulties in extrap olating sp ectral models from the infrared/optical/ultraviolet region to the X-rays, the former of which is typically a mixture of jet and disc emission. The sp ectral break b etween flat-sp ectrum and optically thin emission is p otentially hidden by the disc comp onent and so cannot necessarily b e observed directly (although see also Russell et al. 2006, Migliari 2008). Instead Hynes et al. (2006) showed that the short-term variability can help to distinguish the comp onents. Alternatively Brocksopp et al. (2006) showed that the long-term variability and (anti)correlation of multiwavelength datasets can also b e crucial to disentangling the contributions from disc, jet and/or corona and we adopt a similar approach here. 1.1 XTE J1118+480

PCA (cts/s) 56.12 ± 0.36 57.07 ± 0.30 55.87 ± 0.17 53.80 ± 0.24 51.14 ± 0.16 47.92 ± 0.22 44.68 ± 0.15 45.22 ± 0.21 42.32 ± 0.14 42.79 ± 0.16 39.53 ± 0.14 37.79 ± 0.28 38.48 ± 0.19 36.99 ± 0.26 35.41 ± 0.20 31.89 ± 0.20 28.55 ± 0.23 29.32 ± 0.23 24.54 ± 0.24 20.78 ± 0.09 19.07 ± 0.16 18.71 ± 0.15 16.47 ± 0.12 15.29 ± 0.17 14.11 ± 0.12 12.32 ± 0.16 8.15 ± 0.17 8.35 ± 0.09 7.95 ± 0.08 6.33 ± 0.16 4.64 ± 0.17 4.54 ± 0.11 4.07 ± 0.14 1.90 ± 0.15

HEXTE (cte/s) 5. 5. 6. 6. 6. 5. 4. 5. 5. 5. 5. 4. 3. 3. 3. 4. 2. 2. 2. 2. 1. 1. 1. 2. 1. 1. 0. 0. 0. 0. 0. 0. 0. 0. 42 56 65 02 00 59 97 05 15 06 18 36 87 69 68 19 75 85 46 41 84 05 74 23 25 42 99 57 85 41 80 54 04 13 ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 66 62 21 34 23 31 23 33 22 23 21 84 32 43 34 37 39 41 43 17 35 31 25 42 27 39 61 28 28 42 45 29 39 41

XTE J1118+480 was discovered in 2000 when it entered its first known outburst (Remillard et al. 2000). The event lasted 7 months, decaying after a first p eak and then rebrightening to a "plateau state" for the latter 5 months, staying in the low/hard state throughout (Chaty et al. 2003; Brocksopp, Bandyopadhyay & Fender 2004). While a synchrotron sp ectrum at hard X-rays remains controversial, it was indep endently considered likely in the infrared, optical and ultraviolet regions (Kanbach et al. 2001; Hynes et al. 2003). Frontera et al. (2003) showed that the X-ray sp ectra could b e describ ed by a model of thermal Comptonization plus blackb ody, but that the temp oral prop erties were indicative of a non-thermal, and likely synchrotron, comp onent to the soft X-rays. More recently Reis, Miller & Fabian (2009) discovered a thermal disc comp onent with temp erature 0.21 keV in Chandra sp ectra (although see also Gierlinski, Done & Page 2008, 2009). We note that all these ´

observations took place during the second, plateau-like, phase of the outburst with the earlier p eak only b eing discovered retrosp ectively in RXTE/ASM and CGRO/BATSE data (Remillard et al. 2000; Wilson & McCollough 2000). The second known outburst of XTE J1118+480 b egan in 2005 January, discovered at optical wavelengths (Zurita et al. 2005) and confirmed by X-ray and radio observations (Remillard et al. 2005; Pooley 2005). RXTE observations confirmed that the X-ray source again remained in the low/hard state (Swank & Markwardt 2005; Zurita et al. 2006). A further reflare occurred in the optical, sup erc ?? RAS, MNRAS 000, ??­??


Disentangling jet and disc emission from the 2005 outburst of XTE J1118+480
Table 2. LT photometry in B and V bands and UKIRT photometry in J , H and K bands. Values are given magnitudes. MJD 53393. 53393. 53395. 53395. 53396. 53396. MJD 53390. 53390. 53391. 53391. 53393. 53393. 53394. 53394. 53395. 53395. 53409. 53412. 53426. 4 6 4 6 4 7 4 7 4 6 6 5 5 13. 13. 13. 13. 13. 13. 13. 13. 13. 13. 14. 15. 17. 0 1 0 1 0 1 J 08± 05± 05± 14± 28± 20± 29± 27± 39± 34± 80± 93± 06± 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 01 01 01 01 01 01 01 01 01 01 00 01 03 12. 12. 12. 12. 12. 12. 12. 12. 13. 13. 14. 15. 16. 13. 13. 13. 13. 13. 13. B 64 60 86 90 98 91 ± ± ± ± ± ± 0. 0. 0. 0. 0. 0. 06 04 06 05 06 05 H 66± 67± 72± 80± 96± 91± 95± 97± 04± 04± 56± 40± 24± 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 02 02 02 02 02 02 02 02 02 02 01 01 03 12. 12. 12. 12. 12. 12. 12. 12. 12. 12. 14. 15. 15. 13. 13. 13. 13. 13. V 54 61 60 73 72 ­ ± ± ± ± ± 0. 0. 0. 0. 0. 04 05 04 05 05 K 17± 11± 28± 24± 40± 51± 55± 54± 60± 65± 11± 00± 97± 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 02 02 02 02 02 02 02 02 02 02 01 01 04

3

(High Energy X-ray Timing Exp eriment) instrument onb oard the Rossi X-ray Timing Explorer (RXTE). Data from b oth clusters were added and then binned so as to achieve a minimum of 200 counts/bin using grppha. We note that, while detector 2 on Cluster 1 lost its sp ectral capabilities during the early stages of the mission, this is accounted for during the standard data reduction. Furthermore we have checked the sp ectra of the plus and minus background p ositions (for more details see e.g. Section 3 of Rothschild & Lingenfelter 2003) for our observations and confirmed that there was no noticeable difference b etween the clusters. Soft X-ray observations were obtained from the public archives of b oth the All Sky Monitor (ASM; 2­12 keV) and the Prop ortional Counting Array (PCA; 2­20 keV) onb oard RXTE, the latter of which gives b etter sensitivity but less frequent observations. We extracted sp ectra from the PCA data using ftools version 6.1.1, including data only from the Prop ortional Counter Unit 2 which was functioning throughout. We used the background model appropriate for faint sources and corrected the sp ectra for dead-time (although the source count-rate is such that dead-time effects are small). We added a systematic error of 0.6% to the count-rate in each sp ectral bin using grppha. The resultant PCA count rates are listed in Table 1. 2.2 Optical / Infrared

imp osed on the decay profile, (Chou et al. 2005) but not the radio (Rup en, Dhawan & Mioduszewski 2005). Hynes et al. (2006) studied the short-term variability of the infrared, optical and X-ray emission and discovered an optically thin synchrotron origin to the variable infrared comp onent. A similar result was discovered from comparison of the optical and infrared SEDs (Zurita et al. 2006).

2

OBSERVATIONS

Observations took place at hard and soft X-rays, optical, infrared and radio with a view to obtaining full sp ectral coverage for a large numb er of ep ochs, as required for monitoring of the sp ectrum as the outburst evolved.

B - and V -band observations were obtained via a Target of Opp ortunity prop osal at the Liverp ool Telescop e (LT). The RATCam instrument was used and the data were reduced using the standard LT pip eline routines1 . The resultant apparent magnitudes were extracted using Gaia, with reference to observations of standard stars, and are listed in Table 2. Additional optical photometry has b een retrieved from the literature (Zurita et al. 2005a,b,c,2006; Chou et al. 2005a,b). J -, H - and K -band observations were obtained via a Target of Opp ortunity prop osal at the UK Infrared Telescop e (UKIRT). The UFTI (UKIRT Fast Track Imager) instrument was used and the data were reduced using the standard UKIRT software, oracdr, again with reference to standard stars. The resultant apparent magnitudes are given in Table 2. Additional near-infrared photometry has b een extracted from the literature (Hynes et al. 2005, 2006) 2.3 Radio

2.1

X-rays

Hard X-ray observations in the energy range 14­195 keV were obtained by the Burst Alert Telescop e (BAT) on-b oard Swift. These observations were particularly fortuitous since Swift had b een launched less than two months prior to the outburst of XTE J1118+480. The BAT had only b een collecting data for a month, making this outburst one of the first transient events observed by Swift. The other Swift instruments were not yet in full op erational mode and no p ointed observations were achievable. However, the large field of view of the BAT meant that XTE J1118+480 was detected on ten occasions while still bright. The sp ectra were processed in a manner similar to that for GRO J1655-40 (see Brocksopp et al. 2006 for full details) and the count rates listed in Table 1. Further hard X-ray observations in the range 20­200 keV were obtained using the public archive of the HEXTE
c ?? RAS, MNRAS 000, ??­??

We extracted VLA radio observations at 1.42, 4.7, 8.4, 15 and 23 GHz from the public archive of the Very Large Array (VLA). The array was in the A configuration at the b eginning of the outburst, switching to the AB configuration on 2005 January 14 and the source was detected during 10 ep ochs. The flux calibrators 3C 286, 3C 147 and/or 3C 48 were used dep ending on time and wavelength. Phase referencing was applied using the calibrators 1146+539 (JVAS J1146+5356), 11534+49311 (JVAS J1153+4931) or 11270+45161 (JVAS J1126+4516), dep ending on wavelength. The data were reduced using standard flagging, calibration and imaging routines within the National Radio Ob1

http://telescope.livjm.ac.uk/Info/TelInst/Pipelines/


4

Brocksopp et al.

"shoulder"

reflare

Figure 1. Lightcurves for each waveband. Top: Swift/BAT hard X-rays (14­195 keV). Second: RXTE/HEXTE (20­100 keV). Third: RXTE/ASM (2­12 keV; crosses) and RXTE/PCA (3­20 keV; solid symbols). Fourth: optical and infrared points, from this work (solid symbols) and the literature (other symbols). Horizontal line indicates R-band quiescence (Zurita et al. 2006). Error-bars are to within the size of the points. Bottom: radio data from the VLA (solid symbols) and Ryle Telescope (open symbols). Small symbols (after MJD 53410) are 3 upper limits at 4.7 and 8.4 GHz. Inset plot shows more clearly the deviation from a smooth decay at 4.7 GHz. The dashed vertical line in all plots represents the time of the PCA peak. c ?? RAS, MNRAS 000, ??­??


Disentangling jet and disc emission from the 2005 outburst of XTE J1118+480

5

Table 3. VLA flux densities at 1.42, 4.7, 8.4, 15 and 23 GHz. 1 errors are listed for detections; 3 upper limits are included for non-detections. Ryle Telescope daily-averaged flux densities at 15 GHz are also listed. MJD 1.42 GHz 53383. 53383. 53384. 53384. 53385. 53386. 53386. 53387. 53388. 53391. 53393. 53393. 53394. 53396. 53396. 53399. 53401. 53404. 53407. 53414. 53421. 53428. 2 3 0 7 4 0 2 4 4 6 0 9 3 1 5 3 3 2 5 0 0 0 ­ ­ ­ ­ ­ ­ 45± ­ ­ 58± ­ ­ ­ ­ 76± 50± ­ ­ ­ ­ ­ ­ Radio Flux Density (mJy) 4.7 GHz 8.4 GHz 15 GHz ­ 5.85± 0.13 ­ 5.70± 0.20 ­ ­ 4.44± 0.19 ­ ­ 3.97± 0.17 ­ ­ 2.74± 0.18 ­ 2.38± 0.17 1.86± 0.16 ­ 0.40± 0.13 0.21± 0.10 0.13 0.15 -- ­ 12.29± 0.15 ­ 7.69± 0.15 ­ ­ 6.12± 0.13 ­ ­ 5.29± 0.10 ­ ­ 3.35± 0.11 ­ 2.57± 0.10 1.81± 0.10 1.55± 0.12 0.96± 0.14 0.40± 0.13 0.13 0.10 0.12 8.7± 1.9 8.74±0.70 11.2±2.0 ­ 9.6± 2.8 8.0± 2.1 ­ 7.8± 2.3 7.8± 1.5 ­ 3.7± 1.9 4.2± 1.9 ­ 3.4± 2.1 2.69±0.53 2.05±0.42 ­ ­ ­ ­ ­ ­

23 GHz 7.44±0.49 ­

2.

0.30

­

1.

0.27

5.34± 0.32

5.18± 0.30 3.61± 0.28 2.55± 0.19 ­ 1.25± 0.24 ­ ­ ­ ­

0. 0.

0.20 0.22

servatory's Astronomical Image Processing System ( aips). The flux densities of the primary calibrators were obtained using the formulae of Baars et al. (1977) but with the revised coefficients of Rick Perley, as is the default option in the setjy routine. The flux densities of XTE J1118+480 were calculated by fitting a Gaussian to the p osition of the source; each detection showed the source as an unresolved p oint source. Additional radio observations at 15 GHz were obtained at the Ryle Telescop e of the Mullard Radio Astronomy Observatory, Cambridge, UK. Further details of the observing technique can b e found in Pooley & Fender (1997). All VLA and Ryle flux densities are listed in Table 3

flare with p ossible related features at lower frequencies (see b elow). The soft X-rays at 2­12 keV (ASM: crosses) and 3­20 keV (PCA: solid circles), plotted in the third panel, show similar b ehaviour. They rise to a single p eak, b efore decaying smoothly with a similar "shoulder"-like feature sup erimp osed on the PCA decay that coincides loosely with that of the BAT and HEXTE. The apparent rebrightening in the ASM data (MJD 53412) and also in one of the PCU detectors on the PCA (PCU 0 only, which we do not include in Fig. 1) app ears to b e a background event, despite the quasisimultaneity with the features in HEXTE and the optical. The fourth panel shows the optical and infrared data and there is a very different lightcurve morphology; the decay is much slower than that of the X-rays. A second, smaller p eak is observed in the V -band after the source had b een decaying for ab out a month, although this occurs a few days after the reflare of the hard X-rays. We note that the BAT and ASM p oints b efore and after the outburst represent nondetections rather than detected quiescence and, as such, cannot b e used to claim that the optical rise preceded that of the X-rays. Finally the b ottom panel shows the radio data. No p oints were obtained during the rise and so the radio must have p eaked b efore or simultaneously with our first observation on MJD 53383. The radio p eak therefore apparently precedes or coincides with the X-ray and optical p eaks. The uncharacteristic broken decay of the b etter-sampled frequencies (4.7 and 8.4 GHz) hints at an additional ejection event, presumably related to the features seen in the X-ray lightcurves. The radio source then decayed to b elow 3 b etween MJD 53407 and 53414.

3

RESULTS - LIGHTCURVES

The resultant lightcurves for the 2005 outburst of XTE J1118+480 are shown in Fig. 1. There is good temp oral coverage in all wavebands and there are many ep ochs with observations obtained simultaneously in more than one sp ectral region. The top two panels of Fig. 1 show the hard X-ray temp oral b ehaviour as observed with BAT and HEXTE. The onset of the outburst was detected by the BAT instrument, showing a steep rise in the numb er of counts from MJD 53380 to a maximum around MJD 53384. HEXTE b egan p ointed observations just prior to the hard X-ray p eak (MJD 53384­5). They then decayed with at least one "shoulder" (e.g. MJD 53392) sup erimp osed on the otherwise exp onential decay profile (there are p ossibly two features on the HEXTE lightcurve ­ MJD 53388 and MJD 53392). The final data-p oint of note is at MJD 53410, an apparent rec ?? RAS, MNRAS 000, ??­??


6

Brocksopp et al.
MJD 53384

10 1 0.1 0.01 10 10 10
-3

normalized counts s-1 keV

RXTE/PCA

RXTE/HEXTE

-4

Swift/BAT

-5

Table 4. Results of fitting a simple power-law spectral model to the combined RXTE/PCA, RXTE/HEXTE and Swift/BAT data simultaneously. Where there is more than one RXTE observation per day, we have used the single BAT spectrum with each RXTE spectrum. We adopt Nh = 1.3 â 1020 and include a Gaussian of width 0.01 keV at 6.6 keV (following Maitra et al. 2009). The columns list values for the photon index and normalisation of the power law, the reduced 2 and number of degrees of freedom. Errors are to within 90% confidence. Epoch 53384a 53388a 53388b 53390a 53390b 53390c 53392a 53392b 53392c 53394a 53394b 53394c 53394d 53396a 53396b 53398a 53398b PL Index 1. 1. 1. 1. 1. 1. 1. 1. 1. 1. 1. 1. 1. 1. 1. 1. 1. 76 77 76 79 78 78 79 79 84 83 83 83 84 84 83 91 83 ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 02 01 01 02 01 02 03 02 02 01 02 03 03 02 04 09 04 PL Normalisation 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 137 101 095 086 094 089 069 071 066 054 044 049 050 039 033 025 023 ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 003 002 002 003 002 002 004 002 003 001 001 002 002 001 002 004 002 0. 0. 1. 1. 0. 0. 0. 0. 0. 0. 1. 0. 0. 0. 0. 0. 1.
2

-1

0.2

residuals

0 -0.2 -0.4 5 10 20 Energy (keV) 50 100

dof 73 196 212 172 185 118 125 119 124 212 176 134 147 176 118 80 163

Swift/BAT

RXTE/PCA

RXTE/HEXTE

75 86 09 13 98 92 90 70 97 89 07 94 82 98 93 85 00

keV2 (Photons cm-2 s-1 keV-1)

0.1

1

5

10

20 Energy (keV)

50

100

Figure 2. The combined RXTE/PCA, RXTE/HEXTE and Swift/BAT data from MJD 53384 (53384a in Table 4) fitted simultaneously with a simple power-law model. The data have been rebinned to 5 for the plot only. Top panel: counts spectrum with residuals, the PCA, HEXTE and BAT points are indicated by crosses, circles and squares respectively. Bottom panel: unfolded energy spectrum. The offset in normalisation between the BAT and HEXTE data is surprising and indicates the need for a careful study of spectral fits with the BAT survey data in relation to other instruments. Such study is beyond the scope of this paper but the discrepancy does not affect the results presented here.

4

RESULTS - SPECTRA

Detailed broadband modelling of data from MJD 53393 can b e found in Maitra et al. (2009). For this study we are interested only in the variability of the overall sp ectra; therefore we fit the soft/hard X-ray sp ectrum with a simple p owerlaw, using the model powerlaw within xspec V11.3.2. This simple model was successful at fitting the hard X-ray sp ectra in the range 14­195 keV, yielding values of in the range 1.5­2.1. Similarly the model was successful at fitting the RXTE/PCA sp ectra in the range 3­20 keV with 1.7 < < 1.9; we also included a Gaussian of width 0.01 keV and energy 6.6 keV, representing emission due to Fe K, which b oth improves the fit and provides consistency with Maitra et al. (2009). The resultant p ower-law fits are consistent with the X-ray source having remained in the low/hard state for the duration of the outburst. Having fit the hard and soft X-rays separately, we then fit Swift/BAT, RXTE/HEXTE and RXTE/PCA simultaneously. The resultant fit parameters are listed in Table 4 and show that this simple p ower-law model provides a good fit to the data.

A sample sp ectrum is plotted in Fig. 2. We see no evidence for an exp onential cut-off to the hard X-ray sp ectrum up to 200 keV using the phenomenological cut-off p owerlaw model, cutoffpl. Detailed broadband (radio to X-ray) sp ectral fits to quasi-simultaneous data at multiple ep ochs during the outburst decay will b e presented in a forthcoming pap er (Maitra et al. 2010, in prep.). The optical and infrared sp ectra have b een analysed in depth by Zurita et al. (2006) and Hynes et al. (2006) and so we do not rep eat their work here. Radio sp ectral indices, (defined by S , where S is the flux density at frequency ) have b een calculated for each ep och for which there were two or more observing frequencies. They are listed in Table 5 and the sp ectra plotted in Fig. 3. In each case, is consistent with the inverted sp ectrum we would exp ect in the low/hard state. However, the high values of 2 reflect the p oor fits of the data to a p owerlaw, particularly for MJD 53383, as can b e seen in Fig. 3. It app ears that although the sp ectra are predominantly partially self-absorb ed, as indicated by the p ositive values of , there is also optically thin emission present at higher frequencies for some ep ochs. The observed radio sp ectrum is the comp osite of optically thin ( < 0 ) radio emission from ejections earlier in the outburst that blend with the core flat-sp ectrum ( =0) radio emission (see e.g. Jonker et al. 2009). Fig. 3 and Table 5 suggest tentatively that the contribution of optically thin emission may b e asscociated with the initial outburst and the radio "shoulder" at MJD 53391, when the p ower-law fits are worst, and decrease as the radio source decays. Again, however, the p oor fits to the sp ectra require us to b e cautious in drawing such a conclusion.
c ?? RAS, MNRAS 000, ??­??


Disentangling jet and disc emission from the 2005 outburst of XTE J1118+480

7

Table 5. Spectral indices for VLA data. The spectral index, , is defined in terms of S , where S is the flux density at frequency . Epochs with two or more frequencies have been included. We note that the values of may be misleading as the total radio emission may incude residual optically thin emission, from the onset of the outburst, at higher frequencies at some epochs. The relatively poor fits to a power-law are reflected in the high values of 2 shown in the fourth column. MJD 53383. 53384. 53386. 53391. 53394. 53396. 53399. 53404. 53407. 3 7 2 6 3 5 3 2 5 0. 0. 0. 0. 0. 0. 0. 0. 0. 39 42 43 31 41 40 34 37 31 err 0. 0. 0. 0. 0. 0. 0. 0. 0. 02 07 05 03 05 05 06 18 94
2

DOF 2 0 1 2 1 3 3 1 0

743.0 1.9 3.4 54.1 0.3 11.9 10.3 4.8 0.7

Figure 3. Radio spectra for those epochs with four or more frequencies. An optically thin turnover at higher frequencies can be seen in the top plot and may be "contaminating" the others. Best-fit straight lines of gradient are plotted on each panel.

5

RESULTS - FLUX:FLUX CORRELATIONS
Figure 4. Superimposed HEXTE, radio and R-band lightcurves; the R-band points have been converted into dereddened flux density for more accurate comparison. There is significant deviation between the R-band and the other lightcurves.

In Section 3 we commented on the degree of flux correlation b etween the various wavebands and investigate it further here. We use the HEXTE data as a b enchmark lightcurve since it has well-sampled data and is least likely to b e contaminated with disc emission. Using E (B - V ) = 0.21 (Hynes et al. 2000), the extinction law of Cardelli, Clayton & Mathis (1989) and the flux conversions of Bessell, Castelli & Plez (1998), we converted the R-band brightness values to extinction-corrected flux density and overplotted the HEXTE, R-band and 4.7-GHz lightcurves (Fig. 4). It is immediately apparent that the X-ray and radio lightcurves have similar slop es, while the R-band deviates during its decay. In order to quantify this correlation, we used the p owerlaw sp ectral fits to determine the HEXTE and PCA flux and plot flux:flux plots for HEXTE:PCA, HEXTE:R-band and HEXTE:radio(4.7-GHz) (Fig. 5). Data-pairs were as simultaneous as p ossible and always to within one day. The resultant Sp earman correlation coefficients, , are 0.95 to within 99% confidence. Such a high correlation is surprising for the R-band and may reflect a change in the sp ectrum (to b e investigated further in a future work) ­ from the lightcurves we susp ect that the correlation would worsen as time progresses. The filled circles indicate extra p oints (not included in the fit) for which HEXTE and PCA sp ectra could not b e obtained due to the low count-rates; approxc ?? RAS, MNRAS 000, ??­??

imate fluxes were derived using WebPimms and assumed photon-index of 1.8. The two p oints which fall off the correlation reflect the HEXTE reflare.

6

DISCUSSION

The lightcurves of the 2005 outburst of XTE J1118+48 are notable for their FRED-like (fast rise, exp onential decay) morphology. They show a high degree of correlation b etween the radio and X-ray emission, suggesting a close relation b etween the emission processes and/or emitting regions (see Zdziarski & Gierlinski (2004) for discussion of how the correlation does not require the X-ray and radio emission to b e produced via the same process). The optical lightcurve decays more gradually than those of the other bands. There also app ears to b e an optically thin comp onent to the radio emission, which p ossibly decays in favour of the selfabsorb ed jet as the outburst proceeds.


8

Brocksopp et al.
Done & Page (2008, 2009) have explained the softening of the sp ectrum in terms of irradiation of the inner disc and Done & D´az Trigo (2009) demonstrated that the iron line i is an artefact of pile-up, thereby resolving the controversy over whether or not the disc is truncated. This is also discussed in Maitra et al. (2009), whose modelling of one ep och during the decay of XTE J1118+480 in 2005 allows for (although does not require) a small inner disc radius and, consequently, a source of soft X-rays with which to irradiate the outer disc. For a typical low/hard state outburst we would exp ect the p ower-law comp onent of the X-ray emission to dominate at b oth high and low energies and for there to b e little difference b etween the resp ective lightcurves. That indeed app ears to b e the case in the high and low energy X-ray lightcurves presented here and there is no suggestion that there are contributions from more than one comp onent (contrasting with the lightcurves of, for example, GRO J1655-40; Brocksopp et al. 2006). Once the X-rays and radio have started to decay, their b ehaviour is closely linked. They all app ear to undergo some additional event around MJD 53385­53390, manifested as a "shoulder" sup erimp osed on the decay. The profile of the radio lightcurve is more comparable with that of sources which made transitions to softer sp ectral states, such as XTE J1720-318 or A0620-00 (Brocksopp et al., 2005; Kuulkers et al. 1999 resp ectively) than we would exp ect for a source in the low/hard state. A second, optically-thin, jet ejection event seems a likely cause of this radio "shoulder", particularly as the radio sp ectrum at this time is particularly p oorly fit by a p ower-law. Confirmation of this would require detection of a simultaneous reduction in the sp ectral index, needing higher sensitivity and sampling in the radio data. Multiple ejections are commonly associated with transient events (Brocksopp et al. 2002) but more usually with the optically thin events of sources which soften and enter the very high state (Fender et al. 2004). Such jet ejections may b e unexp ected during the low/hard state but not unprecedented (e.g. GS 1354-64 Brocksopp et al. 2001). It may b e that the "canonical" stable, compact flat-sp ectrum jet of the low/hard state (e.g. Cyg X-1, GX 339-4 or the 2000 outburst of XTE J1118+480; Markoff et al. 2005 and references therein) may turn out to b e the exception rather than the norm. Alternatively, this "ejection" may b e more like the flare seen in V404 Cyg during quiescence, thought to b e some sort of re-energising of the electrons within the jet rather than a new ejection event (Miller-Jones et al. 2008), alb eit on a much longer timescale. Given the link b etween the p ower-law X-ray emission and the jet it might seem surprising that the reflare at MJD 53415 was detected at optical and hard X-ray wavelengths but not the radio. We note that the X-ray reflare preceded the optical reflare and occurred in a gap b etween radio observations and so we could have missed a similar event in the radio due to the sparse sampling. Multiwavelength daily monitoring of these events is required at high sensitivity to determine their true nature. Finally, we compare these results with the X-ray, optical and radio lightcurves of the 2000 outburst of XTE J1118+480, detailed analyses of which were presented by Chaty et al. (2003) and Brocksopp et al. (2004). The lightcurves of that event are notable for their highly correlated b ehaviour at all frequencies, with the long, plateau-like
c ?? RAS, MNRAS 000, ??­??

Figure 5. Flux:flux plots for HEXTE:PCA, HEXTE:R-band and HEXTE:radio(4.7 GHz). Spearman rank correlation components are shown on each panel; 0.95 to within 99% confidence in each case. The best fit straight lines are plotted over the data, accompanied by their respective equations. The three filled circles are additional (unfitted) points, the faint X-ray fluxes of which were approximated using WebPimms. Two of these three points (filled circles) lie off the correlations, reflecting the HEXTE reflare.

The radio emission can b e attributed to a jet on account of its self-absorb ed sp ectrum (Fender 2006 and references therein). If the jet were to dominate the optical and infrared emission then we would exp ect to see a greater correlation b etween their resp ective lightcurves. Indeed, the lightcurves shown here could lead us to assume that the optical and infared are dominated by disc emission were it not for the additional synchrotron comp onents found by Zurita et al. (2006), Hynes et al. (2006) and Maitra et al. (2009). Nonetheless, despite the known additional synchrotron comp onent, the excess of optical/infrared emission following the X-ray/radio decay shows that the disc emission is still likely to b e significant at these wavelengths. There has b een recent debate regarding whether there can b e a disc comp onent to the X-ray sp ectrum when a source is in the low/hard state (e.g. Reis et al. 2009; Rykoff et al. 2007; Miller et al. 2006; Chiang et al. 2009). Gierlinski, ´


Disentangling jet and disc emission from the 2005 outburst of XTE J1118+480
second p eak seen simultaneouly at hard X-ray, soft X-ray, optical and radio wavelengths. Sp ectral fits showed that the synchrotron jet was found to extend to high frequencies, p ossibly the hard X-rays, and dominate the contribution from any disk comp onent (Markoff et al. 2001). In contrast, the 2005 event shows a much more "canonical" FRED lightcurve morphology, often seen in soft X-ray transient events. The optical emission shows a much slower decay than the Xray and radio, suggesting that it is dominated by an alternative comp onent, most likely an accretion disc, although there also app ears to b e a synchrotron comp onent (as discussed ab ove). The flat sp ectrum of the synchrotron emission is contaminated by optically thin radio emission. Different phenomena are clearly dominating the 2000 and 2005 events, despite b oth b eing governed by the prop erties of the low/hard state. Obviously these phenomena cannot b e properties which would remain unchanged over the intervening 5 years, such as orbital parameters, mass of the comp onents or the spin of the black hole. We may need to consider properties intrinsic to the accretion disc or jets instead.

9

by Liverp ool John Moores University in the Spanish Observatorio del Roque de los Muchachos of the Instituto de Astrofisica de Canarias with financial supp ort from the UK Science and Technology Facilities Council. The United Kingdom Infrared Telescop e is op erated by the Joint Astronomy Centre on b ehalf of the Science and Technology Facilities Council of the U.K. The VLA is a facility of the National Radio Astronomy Observatory, which is op erated by Associated Universities Inc., under coop erative agreement with the National Science Foundation.

REFERENCES
B B B B essell M.S., Castelli F., Plez B., 1998, A&A, 333, 231 rocksopp C. et al., 2002, MNRAS, 331, 765 rocksopp C. et al., 2006, MNRAS, 365, 1203 rocksopp C., Bandyopadhyay R.M., Fender R.P., 2004, NewA, 9, 249 Brocksopp C., Corbel S., Fender R.P., Rupen M., Sault R., Tingay S.J., Hannikainen D., O'Brien K., 2005, MNRAS, 356, 125 Brocksopp C., Jonker P.G., Fender R.P., Groot P.J., van der Klis M., Tingay S., 2001, MNRAS, 323, 517 Cardelli J.A., Clayton G.C., Mathis J., 1989, ApJ, 345, 245 Chaty S., Haswell C.A., Malzac J., Hynes R.I., Shrader C.R., Cui W., 2003, MNRAS, 346, 689 Chiang C.Y., Done C., Still. M., Godet O., 2009, MNRAS submitted Chou Y., Chen A., Chiang P.S., Chen A.C., Lin Z.Y., Chung Y., 2005a, ATel. 399 Chou Y., Chen A., Chiang P.S., Hu C.P., Hu R.H., 2005b, ATel. 412 Done C., D´az Trigo M., 2009, MNRAS submitted i Fender R.P., 2006, in Lewin W., van der Klis M., eds. Compact Stellar X-ray Sources, Cambridge Univ Press, Cambridge, ch. 9 Fender R.P., Belloni T., Gallo E., 2004, MNRAS, 355, 1105 Frontera F., Amati L., Zdziarski A.A., Belloni T., Del Sordo S., Masetti N., Orlandini M., Palazzi E., 2003, ApJ, 592, 1110 Gierlinski M., Done C., Page K, 2008, MNRAS, 388, 753 ´ Gierlinski M., Done C., Page K, 2009, MNRAS, 392, 1106 ´ Hynes R.I., Mauche C.W., Haswell C A., Shrader C.R., Cui W., Chaty S., 2000, MNRAS, 539, L37 Hynes R.I., et al., 2003, MNRAS, 345, 292 Hynes R.I. et al., 2006, ApJ, 651, 401 Hynes R.I., Gelino D.M., Pearson K.J., Robinson E.L., 2005, ATel. 393 Jonker P.G. et al., 2009, MNRAS in press Kanbach G., Straubmeier C., Spruit H.C., Belloni T., 2001, Nature, 414, 180 Maitra D., Markoff S., Brocksopp C., Noble M., Nowak M., Wilms J., 2009, MNRAS, 398, 1638 Markoff S., Falcke H., Fender R.P., 2001, A&A, 372, L25 Markoff S., Nowak M.A., Wilms J., 2005, ApJ, 635, 1203 McClintock J., Remillard R., 2006, in Lewin W., van der Klis M., eds. Compact Stellar X-ray Sources, Cambridge Univ Press, Cambridge, ch. 4 Migliari S., 2008, in "A Population Explosion: The Nature and Evolution of X-ray Binaries in Diverse Environments", AIP Conference Proceedings, 1010, 23 Miller J.M, Homan J., Steeghs D., Rupen M., Hunstead R.W., Wijnands R., Charles P.A., Fabian A.C., 2006, ApJ, 653, 525 Miller-Jones J.C.A., Gallo E., Rupen M.P., Mioduszewski A.J., Brisken W., Fender R.P., Jonker P.G., Maccarone T.J., 2008, MNRAS, 388, 1751 Pooley G.G., 2005, ATel. 385 Pooley G.G., Fender R.P, 1997, MNRAS, 292, 925

7

CONCLUSIONS

XTE J1118+480 has now b een observed during two outbursts. We have obtained data for the 2005 event, covering all wavebands at multiple ep ochs. We use analysis of the lightcurves to disentangle jet and disc emission with a view to full sp ectral modelling in a future work. The results p oint towards a very different nature of XTE J1118+480 in 2005 compared with 2000, despite b oth events remaining in the low/hard state and showing correlated X-ray and radio b ehaviour. In 2000 the lightcurves at X-ray, optical and radio wavelengths were characterised by a long, plateau-like phase, which could b e attributed to a dominant and longlasting stable jet, with a flat sp ectrum extending at least through the lower frequencies. In contrast, the 2005 outburst b ehaved more like a canonical soft X-ray transient, with short-lived jet ejection events, an optically thin contribution to the synchrotron sp ectrum, a more dominant disc comp onent and a "fast rise, exp onential decay" lightcurve morphology. These results add to the discussion that the low/hard state covers a wider range of prop erties than typically assumed. Furthermore, they show that the nature of an outburst in the low/hard state must b e governed by at least one prop erty that can vary on a timescale of just a few years.

ACKNOWLEDGMENTS We thank Sandy Leggett for help with the UKIRT data reduction, J¨rn Wilms for advice regarding the HEXTE analo ysis and the anonymous referee for useful suggestions which improved the pap er. PGJ acknowledges supp ort from a VIDI grant from the Netherlands Organisation for Scientific Research. HAK is supp orted by the Swift pro ject. This research has made use of data obtained through the High Energy Astrophysics Science Archive Research Center Online Service, provided by the NASA/Goddard Space Flight Center. The Liverp ool Telescop e is op erated on the island of La Palma
c ?? RAS, MNRAS 000, ??­??


10

Brocksopp et al.

Reis R.C., Miller J.M., Fabian A.C., 2009, MNRAS in press Remillard R., Garcia M., Torres M.A.P., Steeghs D., 2005, ATel. 384 Remillard R., Morgan E., Smith D., Smith E., 2000, IAUC 7389 Rothschild R.E., Lingenfelter R.E., 2003, ApJ, 582, 257 Russell D.M., Fender R.P., Hynes R.I., Brocksopp C., Homan J., Jonker P.G., Buxton M.M., 2006, MNRAS, 371, 1334 Rupen M., Dhawan V., Mioduszewski A., 2005, ATel. 420 Rykoff E.S., Miller J.M., Steeghs D., Torres M.A.P., 2007, ApJ, 666, 1129 Swank J., Markwardt C., 2005, ATel. 390 Wilson C.A., McCollough M.L., 2000, IAUC 7390 Zdziarski A.A., Gierlinski M., 2003, Proceedings of "Stellar-mass, ´ intermediate-mass, and supermassive black holes", Kyoto, 10.2003, Progress of Theoretical Physics Supplement, 155, 99 (astro-ph/0403683 v3) Zdziarski A.A., Lubinski P., Gilfanov M., Revnivtsev M., 2003, ´ MNRAS, 342, 355 Zurita C. et al., 2005a, ATel. 383 Zurita C., Rodr´guez-Gil P., Munoz-Darias T., Torres M.A.P., i ~ Shahbaz T., 2005b, ATel. 402 Zurita C. et al., 2005c, ATel. 424 Zurita C. et al., 2006, ApJ, 644, 432

c ?? RAS, MNRAS 000, ??­??