VLA Observations of XTE J1908+094

Michael Rupen, Vivek Dhawan, & Amy Mioduszewski
(NRAO/Socorro)

Last update: 21 March 2003


Please do not use these results without first contacting M. Rupen. These results are preliminary, and provided here primarily to aid other observers.


All error bars are 1sigma unless otherwise noted.


Background: IAU Circulars &c.

IAUC 7856: 21mar02
P. M. Woods, C. Kouveliotou, M. H. Finger, and E. Gogus, National Space Science and Technology Center; and J. Swank, C. Markwardt, and T. Strohmayer, Goddard Space Flight Center, report the serendipitous discovery of a new x-ray transient, XTE J1908+094, in RXTE PCA observations of the soft-gamma-ray repeater SGR 1900+14, triggered following the burst activity on Feb. 17-18 (GCN 1253). These observations failed to detect the 5.2-s SGR pulsations, pointing towards a possible new source as the origin of the high x-ray flux. An RXTE PCA scan of the region around SGR 1900+14 on Feb. 21 was consistent with emission only from known sources (and no new sources). However, the scans required SGR 1900+14 to be 20 times brighter than its quiescent flux level (GCN 1256). A Director's Discretionary Time Chandra observation on Mar. 11 showed that the SGR was quiescent and did not reveal any new source within the Chandra ACIS field-of-view. A subsequent RXTE PCA scan on Mar. 17, taken in combination with the first scan, required that a new source be included in the fit. The best-fit position is R.A. = 19h08m50s, Decl. = +9 22'.5 (equinox J2000.0; estimated 2' systematic error radius), or approximately 24' away from the SGR source. The source spectrum (2-30 keV) can be best fit with a power-law function including photoelectric absorption (column density N_h = 2.3 x 10**22, photon index = 1.55). Iron line emission is present, but may be due to the Galactic ridge. Between Feb. 19 and Mar. 17, the source flux (2-10 keV) has risen from 26 to 64 mCrab. The power spectrum is flat between 1 mHz and 0.1 Hz, falling approximately as 1/f**0.5 up to 1 Hz. At 1 Hz is seen a broad quasiperiodic oscillation peak and a break to a 1/f^2 power law, which continues to 4 Hz. The fractional rms amplitude from 1 mHz to 4 Hz is 43 percent. No coherent pulsations are seen between 0.001 and 1024 Hz. The authors conclude that XTE J1908+094 is a new blackhole candidate.


IAUC 7861: 27mar02
M. Feroci, Istituto di Astrofisica Spaziale e Fisica Cosmica, CNR, Rome; and L. Reboa, BeppoSAX Science Operation Center, Telespazio, Rome, on behalf of the BeppoSAX team, report: "As a part of the follow-up program on soft gamma-ray repeaters, BeppoSAX serendipitously observed the new x-ray transient XTE J1908+094 (IAUC 7856). Due to technical problems with the battery system of the spacecraft, the source was observed only with the Phoswich Detection System (PDS) in the nominal energy range 15-300 keV. The observation started on Mar. 9.386 UT and lasted until Mar. 12.210, for a net observing time of about 48 000 s. Attributing all the counts detected in the collimated field-of-view of the PDS (1.3 deg, full width at half maximum) to the new source, we can derive a high-energy spectrum for the new x-ray transient. The source is well detected up to > 250 keV, with a net count rate of (18.24 +/- 0.04) counts/s at 15-300 keV. The hard x-ray spectrum is not consistent with an extension of the absorbed power law (photon index 1.55) reported for the energy range 2-30 keV (IAUC 7856). Instead, our data require the introduction of a high-energy cut-off at about 100 keV (but still resulting in an unacceptable fit, with chi**2 of about 5 per degree of freedom). Based on this model, taking into account the 24' off-axis location of the source, we derive a source flux of about 2.9 x 10**-9 erg cm**-2 s**-1 in the energy range 15-100 keV. The BeppoSAX/PDS detection of the source up to 250 keV strongly supports the interpretation of the new x-ray transient as a new blackhole candidate."


IAUC 7873: 10apr02
J. J. M. in 't Zand, Astronomical Institute, Utrecht University, and Space Research Organization Netherlands; and M. Capalbi and M. Perri, ASI Science Data Center, Frascati, report on behalf of a larger team that two target-of-opportunity observations were carried out with BeppoSAX on the proposed black-hole x-ray transient XTE J1908+094 (Woods et al., IAUC 7856; Feroci et al., IAUC 7861): "The second observation was with the Medium-Energy Concentrator Spectrometer on Apr. 2.4-3.8 UT and resulted in an order-of-magnitude improvement of the positional accuracy. We localize the source at R.A. = 19h08m52s.6, Decl. = +9d23'07" (equinox 2000.0; uncertainty 17", 90-percent confidence). This is 0'.9 from the position determined by Woods et al. but consistent within its 2' accuracy."


IAUC 7874: 12apr02
M. P. Rupen, V. Dhawan and A. J. Mioduszewski, National Radio Astronomy Observatory, report: "We have detected a possible radio counterpart to the black-hole candidate XTE J1908+094. Observations with the Very Large Array on Mar. 21 and 22 show a radio source (flux 1.1 +/- 0.1 mJy at 4860 MHz, 0.84 +/- 0.08 mJy at 8460 MHz) at R.A. = 19h08m53s.077, Decl. = +9d23'04".90 (J2000.0, uncertainty about 0".1 in each coordinate), within the 17" error box reported from BeppoSAX (IAUC 7873) and about 8" from its center. The radio spectrum is consistent with this identification, although no variability has yet been seen, and there is no obvious extension in the images (resolution 0".20). Further monitoring is underway."


ATel #86: 13apr02
Search for an Optical Counterpart of XTE J1908+094
R. M. Wagner (LBTO) and S. Starrfield (ASU)
We obtained R-band imagery of the X-ray error box of the black hole candidate XTE J1908+094 (IAUC #7856, #7861, and #7873) on 2002 April 11.44 UT with the Hiltner 2.4-m telescope and 8K CCD camera at the MDM Observatory on Kitt Peak. At the location of the possible radio counterpart reported by Rupen, Dhawn, and Mioduszewski (IAUC #7874), we do not find an optical counterpart brighter than R ~ 23. A R ~ 20 mag star (RA = 09:08:52.75, DEC = +09:23:03.9, J2000) lies 4.8 arcsec west and 1.0 arcsec south of the radio position. Further observations will be useful, particularly at infrared wavelengths.


IAUC 7877: 14apr02
P. Garnavich and J. Quinn, University of Notre Dame; and P. Callanan, University College, Cork, report that images (exposures 1500 s in $I$, 2100 s in $R$), obtained with the 1.8-m Vatican Advanced Technology Telescope on Apr. 8 and 9 UT, show no obvious new sources and no strongly varying sources within the stated position error circles of this transient ({\it IAUC} 7856, 7873) when compared with the red Digitized Sky Survey. No sources (limiting mag I = 22.0; astrometric accuracy 0".2 via six USNO-A2.0 stars) are found within 3" of the radio position (IAUC 7874).


IAUC 7897: 9may02
S. Chaty, Open University; and R. P. Mignani, European Southern Observatory (ESO), Garching, report on the detection in the near-infrared of a likely counterpart to the blackhole candidate XTE J1908+094 (IAUC 7856). Target-of-opportunity observations performed by L. Vanzi (ESO) with the ESO New Technology Telescope on Apr. 25 and 29 show an infrared source that is coincident (within an uncertainty of +/- 0".17 in each coordinate) with the position of the radio counterpart (IAUC 7874). The magnitudes of the source on Apr. 25 were J = 18.7 +/- 0.3, H = 17.2 +/- 0.3, K_s = 16.4 +/- 0.2. Although the source was also visible on Apr. 29, the data do not allow an accurate determination of the magnitude and therefore prevent any variability measurement. Taking the value of column density given on IAUC 7856, and assuming that the source is located near the Galactic center, absolute magnitudes of J = 0.4, H = 0.3, and K = 0.3 are derived. Chaty and Mignani add: "If the source is indeed the actual counterpart of XTE J1908+094, its infrared flux could be significantly contaminated by the accretion disk. In this case, the derived upper limits show that the source is a low-mass system with a companion star of spectral type later than A. Further photometric observations are encouraged to search for variability of this source."


IAUC 8029: 9dec02
M. P. Rupen, A. J. Mioduszewski, and V. Dhawan, National Radio Astronomy Observatory, report the detection of renewed radio emission associated with the x-ray transient and black-hole candidate XTE J1908+094 (IAUC 7856): "After the initial radio detection (IAUC 7874) on Mar. 21, observations at irregular intervals with the Very Large Array at 8.46 GHz gave detections through May 13, then non-detections between June 24 and Nov. 11 (to rms noise levels of 0.04-0.08 mJy/beam). On Dec. 7.99 UT, we again obtained a clear detection of 0.34 +/- 0.055 mJy, at the position of the transient. Observations at other wavelengths, particularly x-ray and near-infrared (cf. IAUC 7897), are urgently requested."


Broad-band X-ray measurements of the black hole candidate XTE J1908+094: in't Zand, J.J.M., Miller, J.M., Oosterbroek, T., & Parmar, A.N. 2002, A&A, 394, 553-560.
Abstract: XTE J1908+094 is an X-ray transient that went into outburst in February 2002. After two months it reached a 2-250 keV peak flux of 1 to 2*E-8 erg cm-2 s-1. Circumstantial evidence points to an accreting galactic black hole as the origin of the X-radiation: pulsations nor thermonuclear flashes were detected that would identify a neutron star and the spectrum was unusually hard for a neutron star at the outburst onset. We report on BeppoSAX and RXTE All Sky Monitor observations of the broad-band spectrum of XTE J1908+094. The spectrum is consistent with a model consisting of a Comptonization component by a ~ 40 keV plasma, a multicolor accretion disk blackbody component with a temperature just below 1 keV and a broad emission line at about 6 keV. The spectrum is heavily absorbed by cold interstellar matter with an equivalent hydrogen column density of 2.5*E22 cm-2, which makes it difficult to study the black body component in detail. The black body component exhibits strong evolution about 6 weeks into the outburst. Two weeks later this is followed by a swift decay of the power law component. The broadness of the 6 keV feature may be due to relativistic broadening or Compton scattering of a narrow Fe-K line.

Radio Light Curves

The figure below shows the evolving radio flux density of XTE J1908+094. The open circles represent the peak flux density; the solid dot show the integrated flux density over a small region including (roughly) a half-beam "guard radius" around the regions of clearly-real emission. In several observations the two overlap and you can only see a solid dot. The upper limits correspond to 3-sigma.

Flux evolution (1may02)

On 23apr02 and 1may02 the 8.5 GHz observations resolved the source into a double; by 4may02, the southern component had faded dramatically (see below). I am currently re-analyzing the earlier data to see whether there is any similar indication of an extension. Unfortunately those data were taken in a narrower bandwidth mode to allow imaging the large initial X-ray error circle, resulting in lower sensitivity. We are continuing to observe the source every few days, but will soon (6may02) begin to lose spatial resolution as the VLA antennas are moved to the next-smaller configuration (BnA).

I have not yet reduced the 1.5 and 15 GHz data.

First images showing a double!

Observations at 8.46 GHz on 23apr02 and 1may02 show a clear double source, with the northern component within 0.15 arcsec of the original point source, and the southern component roughly 0.4arcsec to the south. By 4may02 the southern source seems to have faded (by 2sigma), though it is marginally present (peaking at 0.073+/-0.022 mJy/beam, vs. 1may02 0.156+/-0.034 mJy/beam).

8.46 GHz image (23apr02) 8.46 GHz image ( 1may02) 8.46 GHz image ( 4may02) 8.46 GHz image ( 4may02)

I give two versions of the 4may02 image, the first with the same contours as the previous images, the second using the lower noise level of the 4may02 observations, so that one can see some negative contours. In both cases I've superposed some interesting positions. The big circle in the 0.7arcsec radius, 99% confidence level Chandra position from Sandy Patel, centered on     19 08 53.085   +09 23 04.90
The northern-most tiny cross is the original VLA position, from back when the source was a up there at 1-ish mJy. The two other northern crosses are the positions of the northern component from 23apr02 and 1may02; these agree perfectly (within the small errors) with the 4may02 position. The two small crosses to the south represent the southern component as measured on those same days. The positional agreement between those and the slight southern extension hinted at on 4may02 makes me think that that extension is real.

The 4.86 GHz data from 23apr02 are consistent with an extension as seen here, but the larger beam size does not allow us to resolve the two "dots" at that frequency.

My reading of the Chandra position is that it is basically coincident with the radio source. This is partly because an offset would be so intrinsically interesting that believing it would require fairly overwhelming evidence. This is however an argument for a deeper, HRC observation...

As to the southern dot: if it is indeed fading, that's very strong evidence for the entire radio source being associated with the X-ray transient. If it's not, given the rather high noise level of the initial maps, the southern dot could conceivably have been there forever, hiding below the noise floor. My own feeling is that the southern dot has to be real & associated with the X-ray source, i.e., I think we've resolved the radio counter-part.

For Radio Pundits

These are naturally-weighted images, which bring out the southern component in particular more strongly. Using Briggs' robust=0 gives consistent results, but uniformly lower peak flux densities, suggesting that the individual components may be resolved.

The phase reference source for the first (spectral line) observations was 1950+081, 10.3 degrees from the source. On 23apr02 we switched to 1922+155, which is fainter but only 7.0 degrees away, but stuck in a couple 1950+081 scans to check the astrometry. The strongest evidence that the observed shift is not due to this change is that (1) images of X1908+094 referenced to 1950+081 show the northern component at the same position, slightly south of where it used to be, and (2) images of 1950+081 phase-referenced to 1922+155 (or vice versa) show an unresolved point source which is within 9mas (in each coord.) of the known, correct astrometric position.

Positions

The initial images (21/22mar02) gave a J2000 position of
    19 08 53.077   +09 23 04.90,
with an absolute uncertainty of perhaps 0.1arcsec. The three resolved images give positions for the northern and southern components of
    19 08 53.078   +09 23 04.76       19 08 53.081   +09 23 04.44
    19 08 53.080   +09 23 04.80       19 08 53.083   +09 23 04.38
    19 08 53.078   +09 23 04.77
for 23apr02, 1may02, and 4may02, respectively; the southern source is too faint on 4may02 to allow an independent position estimate. These are based on inverted-parabola fits to find the peak, and may be biased for extended sources. There's also the simple statistical uncertainty for sources which are only 4-6sigma; with a beam of 277x233 mas on 1may02, this corresponds to 60-70mas.


Comments?

Please send any questions, comments, or suggestions to Michael Rupen at the e-mail address given below.


Last modified 21 March 2003

mrupen@nrao.edu