Issue |
A&A
Volume 495, Number 1, February III 2009
|
|
---|---|---|
Page(s) | L1 - L4 | |
Section | Letters | |
DOI | https://doi.org/10.1051/0004-6361:200811396 | |
Published online | 20 January 2009 |
LETTER TO THE EDITOR
Phase-resolved spectroscopy of the accreting millisecond X-ray
pulsar SAX J1808.4-3658 during the 2008 outburst![[*]](/icons/foot_motif.gif)
R. Cornelisse1 - P. D'Avanzo2 - T. Muñoz-Darias1 - S. Campana2 - J. Casares1 - P. A. Charles3,4 - D. Steeghs5,6 - G. Israel7 - L. Stella7
1 - Instituto de Astrofisica de Canarias, Calle via Lactea S/N, 3805 La Laguna, Spain
2 - INAF - Osservatorio Astronomico di Brera, via E. Bianchi 46, 23807 Merate, Italy
3 -
South Africa Astronomical Observatory, PO Box 9, Observatory 7935, South Africa
4 -
School of Physics and Astronomy, University of Southampton, Highfield, Southampton SO17 1BJ, UK
5 -
Department of Physics, University of Warwick, Coventry, CV4 7AL, UK
6 -
Harvard-Smithsonian Center for Astrophysics, 60 Garden Street, Cambridge,
MA 02138, USA
7 -
INAF-Osservatorio Astronomico di Roma, via Frascati 33, 00040
Monteporzio Catone (Rome), Italy
Received 21 November 2008 / Accepted 15 December 2008
Abstract
Aims. We obtained phase-resolved spectroscopy of the accreting millisecond X-ray pulsar SAX J1808.4-3658 during its outburst in 2008 to find a signature of the donor star, constrain its radial velocity semi-amplitude (K2), and derive estimates for the pulsar mass.
Methods. Using Doppler images of the Bowen region, we find a significant (8
)
compact spot at a position where the donor star is expected. If this is a signature of the donor star, we measure
km s-1 (1
confidence), which represents a strict lower limit to K2. Also, the Doppler map of He II
shows the characteristic signature of the accretion disc, and there is a hint of enhanced emission that may be a result of tidal distortions in the accretion disc that are expected in very low mass-ratio interacting binaries.
Results. The lower limit on K2 leads to a lower limit on the mass function of f(M1) 0.10
.
Applying the maximum K-correction gives 228 < K2 < 322 km s-1 and a mass ratio of 0.051 < q < 0.072.
Conclusions. Despite the limited S/N of the data, we were able to detect a signature of the donor star in SAX J1808.4-3658, although future observations during a new outburst are still needed to confirm this. If the derived
is correct, the largest uncertainty in determining of the mass of the neutron star in SAX J1808.4-3658 using dynamical studies lies with the poorly known inclination.
Key words: accretion, accretion disks - X-rays: binaries - stars: individual: SAX J1808.4-3658
1 Introduction
Low-mass X-ray binaries (LMXBs) are systems in which a compact object
(a neutron star or a black hole) is accreting matter, via Roche lobe
overflow, from a low-mass (<1 )
star. Some LMXBs exhibit
sporadic outburst activity but for most of their time remain in a
state of low-level activity (White et al. 1984); we refer to
these systems as transients. In April 1998, a coherent 2.49 ms X-ray
pulsation was discovered with the Rossi X-ray Timing Explorer
(RXTE) satellite in the transient SAX J1808.4-3658
(Wijnands & van der Klis 1998; Chakrabarty & Morgan 1998). This was the first
detection of an accreting millisecond X-ray pulsar (AMXP). Seven more
of these systems have been discovered since then, all of which
are transients with orbital periods in the range between 40 min and
4.3 h and spin frequencies from 1.7 to 5.4 ms. These findings
directly confirmed evolutionary models that link the neutron stars of
LMXBs to those of millisecond radio pulsars via the spinning up of
the neutron star due to accretion during their LMXB phase (e.g.
Wijnands 2006).
To date, six episodes of activity have been detected from SAX
J1808.4-3658 with a 2-3 year recurrence cycle. During the 1998
outburst, a detailed analysis of the coherent timing behaviour showed
that the neutron star was in a tight binary system with a 2.01 h
orbital period (Chakrabarty & Morgan 1998; Hartman et al. 2008). The
mass function derived from X-ray data (
)
and the requirement that the companion fills its Roche lobe led to the
conclusion that it must be a rather low-mass star, possibly a brown
dwarf (Chakrabarty & Morgan 1998; Bildsten & Chakrabarty 2001).
![]() |
Figure 1: Average flux-calibrated spectrum of SAX J1808.4-3658. We have labelled the most important features that are present. The strong absorption features around 3940 and 3980 Å are due to diffuse interstellar bands. |
Open with DEXTER |
We present here optical spectroscopy of SAX J1808.4-3658 obtained during the 2008 outburst. One of the key issues in dynamical studies is to measure the radial velocity of the companion star (K2) and use this value to constrain the optical mass function of the system. During quiescence, these goals can be achieved by tracing the absorption features originating in the photosphere of the companion star. However, the intrinsic faintness of the low-mass companion stars in AMXPs makes such an analysis very difficult. To overcome this problem, Steeghs & Casares (2002) have shown that during phases of high mass-accretion rates, Bowen-blend lines emitted by the irradiated face of the companion star can be used. A precise measurement of K2represents the only way to determine the optical mass function of the system and ultimately constrain the neutron star mass.
2 Observations and data reduction
Optical spectroscopic observations of SAX J1808.4-3658 were carried
out on 27 September 2008 with the ESO Very Large Telescope (VLT), using the
1200B grism on FORS1 with a slit width of 0.7 arcsec. We obtained 16 spectra of
360 s integration each, which corresponds to one
orbital period. The seeing during the observations was in the range
.
Image reduction was carried out following standard procedures:
subtraction of an averaged bias frame and division by a normalised flat
frame. The extraction of the spectra was performed with the
ESO-MIDAS software
package. Wavelength and flux calibration of the spectra were achieved
using a helium-argon lamp and by observing spectrophotometric standard
stars. Our final reduced spectra have a wavelength range from 3600-5000 Å,
a dispersion of 0.72 Å pixel-1, and resolution of R=2200.
Cross correlation of the spectral lines and the Doppler
tomograms were obtained using the MOLLY and DOPPLER packages developed
by Tom
Marsh
.
3 Data analysis
We present the average spectrum of SAX J1808.4-3658 in
Fig. 1. The spectrum is dominated by strong Balmer lines
in absorption, which are due to the optically thick disc in the high
state and are a typical signature of a low-to-intermediate inclination
system. The He II 4686 and the Bowen complex (at
4630-4660) are also clearly detected as emission
features, and we have indicated the most important lines.
For other bright LMXBs, narrow components in the Bowen emission have
been reported that are thought to arise on the irradiated surface of
the donor star (e.g. Steeghs & Casares 2002; Casares et al.
2006; Cornelisse et al. 2007; see also Cornelisse et al. 2008, for
an overview), and here we attempt to find similar features in
SAX J1808.4-3658. To calculate the orbital phase for each spectrum,
we used the recent ephemeris by Hartman et al. (2008), but added 0.25
orbital phase to their phase zero so that it represents inferior
conjunction of the secondary. Unfortunately, the individual spectra
do not have high enough S/N to identify the narrow features, so we
must resort to the technique of Doppler tomography (Marsh & Horne
1988). This technique uses all the spectra simultaneously to probe
the structure of the accretion disc and identify compact emission
features from specific locations in the binary system. However, for
this technique to work, an estimate of the systemic velocity of the
system, ,
is crucial. We do want to note that, by changing the
phase zero of the accurate Hartman et al. (2008) ephemeris, any
potential donor star feature in the map must now lie along the
positive y-axis in the Doppler tomogram. Thus finding a significant
feature there will give strong support to an emission site located on
the irradiated donor star.
To find
we started by applying the double-Gaussian technique
of Schneider & Young (1980) to He II
4686. Since the wings
of the emission line should trace the inner-accretion disc, it should
give us an estimate of not only the already known radial velocity of
the compact object (Chakrabarty & Morgan 1998), but also
.
Using a Gaussian band pass with FWHM of 400 km s-1 and
separations between 500 and 1800 km s-1 in steps of 50 km s-1, we find that, between a separation of 1400 and 1600 km s-1, our values for K1 are close to the one obtained by
Chakrabarty & Morgan (1998), while at larger separations we are
reaching the end of the emission line (see Fig. 2). Also in
this range, our fitted orbital phase zero (
)
is close to 0.5,
further suggesting that we are tracing the radial velocities of a
region close to the neutron star, while the systemic velocity is
stable around -50 km s-1. Despite the large errors on our fits,
we do think this test already gives a good first estimate of
around -50 km s-1.
![]() |
Figure 2:
The derived fit parameters of the radial velocity curve of the
He II
|
Open with DEXTER |
Another test to obtain
is to create Doppler maps for
He II
.
Contrary to the Bowen region (which is very complex due
to the presence of many different lines), He II is a single line that
is usually a good tracer of the accretion disc (see e.g. Cornelisse
et al. 2007; Casares et al. 2006). We searched for
between
-160 and 0 km s-1 in steps of 20 km s-1, and all the maps
show the expected accretion disc structure (see Fig. 3).
We must unfortunately conclude that He II is not very sensitive to
and only suggests a range between -160 and 0 km s-1. We
note that the maps are dominated by an emission feature in the
top-left quarter of the map, which we interpret as the gas-stream
impact point. We also note that further downstream there is enhanced
emission, which might be due to matter streaming along the edge of the
disc as was observed, for example in EXO 0748-676 (Pearson et al.
2006). Finally we note that there is some enhanced emission in the
top-right corner, which might be due to strong tidal interaction (see
below).
Cornelisse et al. (2008) have shown that the strongest narrow
component in the Bowen emission is usually N III 4641. Our
next step was therefore to create Doppler maps of the Bowen region for
between 0 and -120 km s-1 (again in steps of
20 km s-1) including only this line. Only when
was between
-80 and -20 km s-1 was a clear spot present and centred on the x=0axis. For this range we estimated the peak value and FWHM of the spot
as a function of
,
and in Fig. 4 we show how the
ratio of these values change. Around
km s-1(1
confidence), the FWHM/peak value reaches a minimum,
suggesting that here the spot is most compact, and this is the value
we adopt for the systemic velocity. Furthermore, we also noted that, in
the range from -35 to -65 km s-1, the velocity centroid of the
spot was stable between Vy=240-260 km s-1.
To decrease the noise present in the Bowen Doppler map, we included the
most important other lines that are most often present in other LMXBs
(N III 4634 and C III
4647/4650), and present this
map in Fig. 3. To estimate the significance of the
compact spot, we measured the standard deviation of the brightness of
the pixels in the background. We find that the central pixels of the
compact spot are 18
above the background, and even 8
above the second most prominent feature in the map, namely the one at
(-150, -200) in the Bowen map of Fig. 3 (for which it is
unclear whether it is real or an artifact of the tomogram). This strongly
suggests that the compact spot is real.
![]() |
Figure 3:
Doppler maps of He II |
Open with DEXTER |
Finally, to optimise our estimate for
,
we created average
spectra in the rest frame of the donor star, changing
in
steps of 2 km s-1 within our error range. We find that
N III
4640 is most pronounced for
km s-1(1
confidence), and adopt this as our final value, and in
Fig. 5 show the final average spectrum in the rest-frame of
the donor star. We do note that this value is lower than the
300 km s-1 obtained by Ramsay et al. (2008) from the same
dataset, since they provide no errors, we cannot tell if this
difference is significant. However, as stated above, the location of
the spot remains within the quoted error range for a range of
assumed
velocities. To search for variability in the Bowen lines as a
function of orbital phase, we created an average spectrum in the
rest-frame of the donor using only spectra taken between orbital
phases 0.25 and 0.75, and another corresponding to phases 0.75
and 1.25. The strength of N III
4640 did not change between
these spectra, but it is unclear if this is real (suggesting that the
inclination is low) or due to the limited S/N of the dataset.
4 Discussion
We have presented phase-resolved spectroscopy of the accreting
millisecond X-ray pulsar SAX J1808.4-3658, and detected a compact
feature in the Doppler map of the Bowen complex. Although the S/N of
the data is limited, this spot is the most stable and significant
feature for a wide range of
velocities. Furthermore, thanks
to the very accurate ephemeris (Hartman et al. 2008), the spot is at a
position where the donor star is expected, and we conclude that it is
real. Therefore, following detections of a donor star signature in
other X-ray binaries (e.g. Steeghs & Casares 2002; Casares et al.
2006; Cornelisse et al. 2007; see Cornelisse et al. 2008, for an
overview), we also identify this feature as being produced on the
irradiated surface of the donor star. We note that in
Fig. 5 most peaks in the Bowen region appear to line up
with known N III and C III lines (e.g. Steeghs & Casares 2002) when
using our derived values for
and
.
Since the donor star surface must have a lower velocity than the
centre of mass, the observed
km s-1(1
confidence) is a lower limit on the true K2 velocity.
However, it still gives us a strict lower limit on the mass function
of f(M) = M1sin3i/(1 + q)2
0.10
,
where q is
the binary mass ratio M2/M1 and i the inclination of the
system. We can further constrain the mass function by applying the
so-called K-correction (Muñoz-Darias et al. 2005) and using the
fact that K1=16.32 km s-1 (Chakrabarty & Morgan 1998). The
largest K-correction possible is when we assume that there is no
accretion disc and almost all radiation is produced in the L1point. Applying the polynomials by Muñoz-Darias et al. (2005) for
km s-1 gives
km s-1,
which should be independent of the inclination of the system.
This gives conservative estimates of 228<K2< 322 km s-1 and 0.051 < q < 0.072, which we used to create the Roche lobe and gas streams on the Bowen Doppler map in Fig. 3. We note that our obtained mass ratio is rather extreme and would suggest that tidal interaction in SAX J1808.4-3658 is important enough to produce a precessing accretion disc and thereby a superhump (see e.g. O'Donoghue & Charles 1996). This might be the explanation of the enhanced emission in the top-right quarter of the He II map (Fig. 3/top), which was for example also observed in LMC X-2 (Cornelisse et al. 2007). Such a superhump should be moving through the accretion disc on the precession time scale, and therefore changes position in the Doppler map over time. Unfortunately, since we only have one orbit of data we cannot test this, and future observations will be needed to see if the spot is long-lived and moves, in order to confirm the presence of the superhump.
![]() |
Figure 4:
The ratio of the FWHM over the peak value for the compact spot
in the Bowen map (using only N III |
Open with DEXTER |
Despite our wide range of K2 we nonetheless will review the
implications for the mass of the neutron star. First of all we can
improve our estimate of K2 by taking the results by Meyer &
Meyer-Hofmeister (1982) into account. They analysed the effects of
X-ray heating on the accretion disc and found that there is a minimal
disc opening angle of 6
,
even in the absence of
irradiation. Using this value to estimate the K-correction, the
polynomials by Muñoz-Darias (2005) suggest that K2
310 km s-1, which is still comparable to the maximum correction
possible. Deloye et al. (2008) used photometry of SAX J1808.4-3658
in quiescence to constrain the inclination between 36 and 67 degrees.
Using these extreme values for their inclination and
228 < K2 < 322 km s-1 leads to a neutron star mass between 0.15 and
1.58
.
Although these values are not very constraining for
the neutron star mass, they favour a mass near the canonical value
rather than a massive neutron star. We also note that our values are
lower than the >1.8
estimated by Deloye et al. (2008, for
a 10% error in the distance estimate), but since this corresponds to
a 1.7
difference (only taking our errors into account), we
conclude that this disagreement is marginal.
5 Conclusions
The observations presented here provide evidence that a signature of the irradiated donor star in SAX J1808.4-3658 is detected. Clearly a similar experiment, but at higher S/N and spectral resolution, must be carried out again during a future outburst to obtain more than a single orbital period of data and resolve the narrow lines. This will not only allow us to unambiguously claim the presence of narrow components in the spectra of SAX J1808.4-3658, but also to measure the rotational broadening of the narrow components to further constrain K2 via the relation in Wade & Horne (2003). With these data we have shown a promising way toward constraining K2, and in combination with more quiescent data, we should be able to better constrain the inclination, thereby obtaining the mass of the neutron star in SAX J1808.4-3658.
![]() |
Figure 5:
Blow-up of the Bowen region for the average spectrum ( top) and
the average in the rest-frame of the donor star ( bottom). The bottom
spectrum clearly shows the narrow components. Indicated are
the most important N III ( |
Open with DEXTER |
Acknowledgements
We cordially thank the director of the European Southern Observatory for granting Director's Discretionary Time (ID 281.D-5060(A). We would like to thank the referee, Craig Heinke,for the careful and helpful comments which have improved this paper. R.C. acknowledges a Ramon y Cajal fellowship (RYC-2007-01046). R.C. acknowledges Katrien Uytterhoeven for useful discussions of different analysis techniques. P.D.A. and S.C. thank S. Covino for useful discussions. D.S. acknowledges an STFC Advanced Fellowship. J.C. acknowledges support by the Spanish MCYT GRANT AYA2007-66887. Partially funded by the Spanish MEC under the Consolider-Ingenio 2010 Program grant CSD2006-00070: First Science with the GTC.
References
- Bildsten, L., & Chakrabarty, D. 2001, ApJ, 557, 292 [NASA ADS] [CrossRef] (In the text)
- Casares, J., Cornelisse, R., Steeghs, D., et al. 2006, MNRAS, 373, 1235 [NASA ADS] [CrossRef] (In the text)
- Chakrabarty, D., & Morgan, E. H. 1998, Nature, 394, 346 [CrossRef] (In the text)
- Cornelisse, R., Steeghs, D., Casares, J., et al. 2007, MNRAS, 380, 1219 [NASA ADS] [CrossRef] (In the text)
- Cornelisse, R., Casares, J., Muñoz-Darias, T., et al. 2008, in A Population Explosion: The Nature and Evolution of X-ray Binaries in Diverse Environments, AIP Conf. Proc., 1010, 148 (In the text)
- Deloye, C. J., Heinke, C. O., Taam, R. E., & Jonker, P. G. 2008, MNRAS, accepted (In the text)
- Hartman, J. M., Patruno, A., Chakrabarty, D., et al. 2008, ApJ, 675, 1468 [NASA ADS] [CrossRef] (In the text)
- Marsh, T. R., & Horne, K. 1988, MNRAS, 235, 269 [NASA ADS] (In the text)
- Meyer, F., & Meyer-Hofmeister, E. 1982, A&A, 106, 34 [NASA ADS] (In the text)
- Muñoz-Darias, T., Casares, J., & Martinez-Pais, I. G. 2005, ApJ, 635, 502 [NASA ADS] [CrossRef] (In the text)
- O'Donoghue, D., & Charles, P. A. 1996, MNRAS, 282, 191 [NASA ADS] (In the text)
- Pearson, K. J., Hynes, R. I., Steeghs, D., et al. 2006, ApJ, 648, 1169 [NASA ADS] [CrossRef] (In the text)
- Ramsay, G., Elebert, P., Callahan, P., Reynolds, M., & Hakala, P. 2008, ATel, 1752 (In the text)
- Schneider, D. P., & Young, P. 1980, ApJ, 238, 946 [NASA ADS] [CrossRef] (In the text)
- Steeghs, D., & Casares, J. 2002, ApJ, 568, 273 [NASA ADS] [CrossRef] (In the text)
- Wade, R. A., & Horne, K. 1988, ApJ, 324, 411 [NASA ADS] [CrossRef]
- White, N. E., Kaluzienski, J. L., & Swank, J. H. 1984, in High Energy Transients in Astrophysics, AIP Conf. Proc., 115, 31 (In the text)
- Wijnands, R. 2006, in Trends in Pulsar Research, ed. J A. Lowry (New York: Nova Science Publ.), 53 (In the text)
- Wijnands, R., & van der Klis, M. 1998, Nature, 394, 344 [CrossRef] (In the text)
Footnotes
- ... outburst
- Based on observations made with ESO Telescopes at the Paranal Observatory under programme ID 281.D-5060(A).
- ...
ESO-MIDAS
- http://www.eso.org/projects/esomidas/
- ...
Marsh
- http://deneb.astro.warwick.ac.uk/phsaap/software/
All Figures
![]() |
Figure 1: Average flux-calibrated spectrum of SAX J1808.4-3658. We have labelled the most important features that are present. The strong absorption features around 3940 and 3980 Å are due to diffuse interstellar bands. |
Open with DEXTER | |
In the text |
![]() |
Figure 2:
The derived fit parameters of the radial velocity curve of the
He II
|
Open with DEXTER | |
In the text |
![]() |
Figure 3:
Doppler maps of He II |
Open with DEXTER | |
In the text |
![]() |
Figure 4:
The ratio of the FWHM over the peak value for the compact spot
in the Bowen map (using only N III |
Open with DEXTER | |
In the text |
![]() |
Figure 5:
Blow-up of the Bowen region for the average spectrum ( top) and
the average in the rest-frame of the donor star ( bottom). The bottom
spectrum clearly shows the narrow components. Indicated are
the most important N III ( |
Open with DEXTER | |
In the text |
Copyright ESO 2009
Current usage metrics show cumulative count of Article Views (full-text article views including HTML views, PDF and ePub downloads, according to the available data) and Abstracts Views on Vision4Press platform.
Data correspond to usage on the plateform after 2015. The current usage metrics is available 48-96 hours after online publication and is updated daily on week days.
Initial download of the metrics may take a while.