THE FIRST HYPERVELOCITY STAR FROM THE LAMOST SURVEY

LAMOST is a 4 m Schmidt telescope (now called the Guo Shoujing Telescope) at the Xinglong Observing Station of the National Astronomical Observatories of China. It is equipped with 4000 optical fibers in the focal plane, taking spectra with resolution R = λ/Δλ = 1800. Within the LAMOST spectroscopic survey (Cui et al. 2012; Zhao et al. 2012), the LAMOST Experiment for Galactic Understanding and Exploration (LEGUE; Deng et al. 2012) aims to take ∼8 million stellar spectra for targets covering 16,000 deg² of sky over the course of five years. The target-selection criteria required to achieve various science goals can be found in Carlin et al. (2012), Chen et al. (2012), Yang et al. (2012), and Zhang et al. (2012).

The internal DR1 of LAMOST spectra and stellar parameters includes spectra obtained during a pilot survey (from 2011 October to 2012 June; Luo et al. 2012) and the first year of the regular survey (from 2012 September). Most of the observed stars are brighter than r = 16. We performed a systematic search for HVSs in the catalog, including stars of A-type or earlier; of these stars, one of the ∼10⁵ stars turned out to be a HVS (hereafter, denoted LAMOST-HVS1).

The star (J091206.52+091621.8; LAMOST-HVS1) has been observed twice by LAMOST, separated by about 70 days (2012 December 23 and 2013 March 5). It is a bright star with magnitude around 13. The radial velocities at the two epochs from spectral fitting are consistent with each other within the uncertainties, and therefore there is no evidence of it being a close binary system.

The measured heliocentric radial velocity, v_r☉ = 620 ± 10 km s⁻¹, translates to a Galactocentric radial component v_rf = 477 ± 10 km s⁻¹, according to

$$\begin{eqnarray} &&v_{{\rm rf}}=v_{r\odot } + U_0\cos l \cos b + (V_{{\rm LSR}}+V_0)\sin l \cos b + W_0\sin b, \nonumber \\ \end{eqnarray} \tag{ 1 }$$

where we adopt V_LSR = 250 km s⁻¹ for the velocity of the local standard of rest (LSR; Reid et al. 2009; McMillan & Binney 2010) and (U₀, V₀, W₀) = (11.1, 12.24, 7.25) km s⁻¹ for the peculiar motion of the Sun with respect to the LSR (Schönrich et al. 2010). The star has proper motion measurements (μ_αcos δ, μ_δ) = (− 4.0 ± 0.7, − 4.9 ± 1.2) mas yr⁻¹ in the UCAC4 catalog (Zacharias et al. 2013) and (− 2.5 ± 1.9, − 1.2 ± 1.9) mas yr⁻¹ in the PPMXL catalog (Roeser et al. 2010). We find that quasars within two degrees around the star in the PPMXL catalog have a net proper motion, and if we correct for this systematic offset, the proper motion of the star becomes (0.9 ± 1.9, 0.9 ± 1.9) mas yr⁻¹ (listed as cPPMXL in Table 1). A more accurate determination is clearly desirable.

Table 1. Properties of LAMOST-HVS1

	J091206.52+091621.8
Position (J2000)	(α, δ) = (138027199, 9272725)
	(l, b) = (221099564, 35407261)
Magnitudes	g = 12.91 r = 13.22 i = 13.50
	B = 12.96 V = 13.06 J = 13.36
	H = 13.43 K = 13.53
Distance	13.4 ± 2.2 kpc (Heliocentric)
	19.4 ± 2.1 kpc (Galactocentric)
Radial velocity	vr☉ = 620 ± 10 km s−1
	vrf = 477 ± 10 km s−1
Proper motion	(μαcos δ, μδ)
(mas yr−1)	(− 4.0 ± 0.7, − 4.9 ± 1.2) [UCAC4]
	(− 2.5 ± 1.9, − 1.2 ± 1.9) [PPMXL]
	(0.9 ± 1.9,     0.9 ± 1.9) [cPPMXL]
Spectral type	B
Teff	(2.07 ± 0.12) × 104 K
log [g/(cm s−2)]	3.67 ± 0.19
[Fe/H]	−0.13 ± 0.07
Mass	9.1 ± 0.7 M☉

Download table as:  ASCIITypeset image

Figure 1 shows the low-resolution LAMOST spectrum of the star. The basic stellar atmospheric parameters, listed in Table 1, are obtained by fitting the spectrum with the University of Lyon Spectroscopic analysis Software (ULySS¹²; Koleva et al. 2009). It has an effective temperature T_eff ≃ 2 × 10⁴ K, surface gravity log [g/(cm s⁻²)] ≃ 3.7, and slightly sub-solar metallicity, [Fe/H] ≃ −0.13. The presence of strong helium lines indicates that it is an early B-type star.

Zoom In Zoom Out Reset image size

To determine the distance and mass of LAMOST-HVS1, we make use of the list of masses and luminosities of B-type stars compiled by Hohle et al. (2010), based on 2MASS photometry and Hipparcos parallax. By matching temperature and surface gravity values to the range derived from spectral fitting of the LAMOST-HVS1, we infer its spectral type to be between B1 and B2.5. While the luminosity class is not well-constrained, it is correlated with the spectral type. For example, the star could be B1I, B2IV, or B2.5V. The degeneracy leads to only a small variation in luminosity among such stars, which translates to a relatively well-constrained distance, 13.4 ± 2.2 kpc. With 8 kpc adopted for the Sun's distance to the GC, the Galactocentric distance of LAMOST-HVS1 is calculated to be R = 19.4 ± 2.1 kpc. The mass of LAMOST-HVS1 is inferred to be 9.1 ± 0.7 M_☉.

Interestingly, LAMOST-HVS1 appears to be almost a twin to HE 0437-5439 (a.k.a HVS 3; e.g., Edelmann et al. 2005; Bonanos et al. 2008; Przybilla et al. 2008), which is also a ∼9 M_☉ B-type star, with similar temperature and surface gravity. The Galactic position of LAMOST-HVS1 shares some similarities with HD 271791 (Heber et al. 2008), a 11 ± 1 M_☉ B-giant star established to be a runaway star with velocity similar to those of HVSs. HD 271791 is 21.8 ± 3.7 kpc away from the GC and −10.4 ± 2.0 kpc below the disk plane (Heber et al. 2008), while for LAMOST-HVS1 those are 19.4 ± 2.1 kpc and 7.8 ± 1.3 kpc. Together, the above three stars make the most massive HVSs discovered so far, and LAMOST-HVS1 is the closest one.

With the velocity and distance determined, Figure 2 places LAMOST-HVS1 in the v_rf–R plane, and compares it to the known HVSs, as well as possible HVSs and possible bound HVSs, as listed in Brown et al. (2012). Clearly, LAMOST-HVS1 is the nearest HVS discovered so far. Following Brown et al. (2012), we also plot two curves (long and short–dashed) of escape velocities, based on the Galactic potential models of Kenyon et al. (2008) and Gnedin et al. (2010), respectively. At R ∼ 20 kpc, the velocity v_rf = 477 km s⁻¹ of LAMOST-HVS1 is above the escape velocity in the Gnedin et al. (2010) model and falls almost on top of the model curve from Kenyon et al. (2008). Note that including the proper motion contribution (using the PPMXL one for the most conservative estimate) implies a total velocity of 545 ± 54 km s⁻¹, with the uncertainties from distance, proper motion, and radial velocity. This establishes LAMOST-HVS1's identity as an HVS.

Zoom In Zoom Out Reset image size

We now explore to what extent the data constrain the origin of LAMOST-HVS1.

Lu et al. (2010) and Zhang et al. (2013) demonstrate that, under the tidal-disruption scenario, the spatial distribution of HVSs can track that of the progenitors. Lu et al. (2010) indeed find that most of the discovered HVSs, if viewed from the GC, show spatial distributions near the great circles connecting to the planes of the clockwise-rotating young stellar (CWS) disk and the northern arm of the mini-spiral or the outer wrapped part of the CWS disk, supporting the GC origin of HVSs (but see also Pawlowski et al. 2013).

Following Lu et al. (2010), the top panel of Figure 3 shows the sky distribution of the known and possible HVSs and the position of LAMOST-HVS1, as viewed from the GC. The great circles correspond to different stellar structures in the GC: the CWS, the outer wrap of the CWS (Outer-CWS), the northern arm (North-arm), and the counter-clockwise stellar disk (CCWS). Interestingly, LAMOST-HVS1 falls into the clustered region defined by most other known HVSs. It is closest to the Outer-CWS great circle and also close to that of the North-arm. This seems to support its GC origin, and suggests that its progenitors were associated with the CWS or North-arm.

Zoom In Zoom Out Reset image size

In the top panel of Figure 3, LAMOST-HVS1's twin, HE 0437-5439, lies near the CCWS circle. However, it is also possible to connect to the Outer-CWS, similar to LAMOST-HVS1. HE 0437-5439 has been proposed to have been produced from the Large Magellanic Cloud (Edelmann et al. 2005; Bonanos et al. 2008; Przybilla et al. 2008), but with well-measured proper motion Brown et al. (2010) conclude that it is more likely a compact binary ejected from the GC, which later evolved into a blue straggler.

The proper motion of LAMOST-HVS1 is not well-measured. We perform a Monte Carlo simulation by accounting for the uncertainties in the proper motion (for the PPMXL, cPPMXL, and the UCAC4 values), distance, and radial velocity (all assumed to be Gaussian), in order to consider the implications of the velocity vector for the origin of LAMOST-HVS1. Obviously, the uncertainties are dominated by those in the proper motion.

From the star's current position, we go along the opposite direction of the velocity vector and perform a full orbit integration with the Galactic potential model in Kenyon et al. (2008) to derive the intercept position in the disk plane. This would be the point of origin for the HVS if it were ejected from the plane. The Monte Carlo simulation shows that the most likely intercept depends on which proper motion measurement is used (see the bottom panel of Figure 3). The GC is within the ∼1σ (<1σ) range of the distribution of the intercept positions with the UCAC4 (PPMXL) proper motion adopted. If we use the PPMXL proper motion corrected for the offset of quasars, the data then favor a disk origin, with the most likely intercept lies close to the Perseus spiral arm. An accurate measurement of the proper motion will be key to determining the origin of LAMOST-HVS1. If the GC origin holds, the flight time (∼32 Myr) for LAMOST-HVS1 to reach its current position would be comparable to its lifetime (estimated to be around 30–40 Myr). If the ejection is delayed by a few million years or more since the formation (e.g., Brown et al. 2012), there would be some tension between the delay + flight time and the lifetime. Therefore, it may be either a massive star ejected directly from the GC (with properties in broad agreements with predictions by Zhang et al. 2013), or a blue straggler from similar processes as HE 0437-5439.

We cannot rule out a disk (runaway) origin of LAMOST-HVS1, especially if we use the cPPMXL proper motion. As mentioned before, the position of LAMOST-HVS1 reminds us of the B-type giant HD 271791, which is about 10 kpc from the disk plane. Its proper motion measurement clearly favors a disk origin, establishing it as a hyper-velocity runaway star (Heber et al. 2008). Bromley et al. (2009) show that high speed stars near the disk should mostly be disk runaway stars, and HD 271791 is therefore an example of unbound runaway stars. For LAMOST-HVS1, a disk runaway origin would greatly alleviate the potential tension between flight time and the lifetime of the star as in the GC-origin scenario. A better proper motion determination for LAMOST-HVS1 will help to show whether the possibility of a disk runaway star holds.

Given that LAMOST-HVS1 is the brightest example of known HVSs and its clear similarity to HE 0437-5439 and HD 271791, it will be extremely interesting to conduct a more detailed study. If indeed it comes from the GC, we expect a proper motion around 3–4.5 mas yr⁻¹, well in reach of current and near future data. High-resolution spectroscopic study will reveal its chemical abundance pattern and rotation velocity. We plan to perform such follow-up observations and investigations.