A New Catalog of Radial Velocity Standard Stars from the APOGEE Data

The (barycentric or heliocentric) stellar radial velocity (RV) of a star is ideally defined as the change rate of distance between the Sun and the star. It can be deduced from the Doppler shift of the spectrum of the star recorded in the reference frame of the telescope, and then transform the derived value to the barycentric or heliocentric reference frame. Accurate stellar RV measurements are of utmost importance for a variety of astrophysical studies, including Galactic kinematics and dynamics (e.g., Bovy et al. 2012, 2015; Huang et al. 2015, 2016, 2017; Sun et al. 2015; Gaia Collaboration et al. 2018b), discoveries of stellar and substellar companions (e.g., binaries, brown dwarfs, and exoplanets; e.g., Mayor & Queloz 1995; Udry & Mayor 2008; Gao et al. 2014, 2017; El-Badry et al. 2018; Tian et al. 2018) and stellar structure and evolution (pulsating variables, binaries and multiple systems; e.g., Chadid et al. 2008; Badenes & Maoz 2012; Yang et al. 2014). The typical accuracies for those various studies are required from few m s⁻¹ (e.g., for discovering of low-mass planets; Udry & Mayor 2008) to few km s⁻¹ (for studying stellar structure and evolution, and Galactic kinematics and dynamics; Gao et al. 2014, 2017; Sun et al. 2015).

In the past decades, the number of stellar RV measurements has increased dramatically, thanks to several either already completed or still ongoing large-scale spectroscopic surveys, e.g., the RAVE (Steinmetz et al. 2006), the SDSS/SEGUE (Yanny et al. 2009), the SDSS/APOGEE (Majewski et al. 2017), the LAMOST (Deng et al. 2012; Liu et al. 2014), the Gaia-ESO (Gilmore et al. 2012), the HERMES-GALAH (De Silva et al. 2015), and the Gaia-RVS (Katz 2009; Katz et al. 2018) surveys. More ambitious large-scale spectroscopic surveys are under-plan and upcoming, such as the WEAVE (Dalton et al. 2014) and the 4MOST (de Jong et al. 2016) surveys.

To achieve accurate RV measurements, it is essential to build sets of RV standard stars suitable for various spectroscopic surveys. High-quality RV standard stars are needed to determine the RV zero points (RVZPs) of the instruments employed by various surveys. As pointed by Crifo et al. (2010), the concept of RV standard stars is based on the physical notion "stability" rather than on other more fundamental physical definitions (e.g., Lindegren & Dravins 2003). As such, the definition of RV standard stars can change depending on the accuracy of RV measurements expected for the surveys.

Up till now, there are a few thousand RV standard stars with velocity uncertainties less than 100 m s⁻¹ over a time baseline longer than one year, owning to efforts of several groups (e.g., Udry et al. 1999, hereafter UMQ99; Nidever et al. 2002, hereafter N02; Chubak et al. 2012, hereafter C12; Soubiran et al. 2013, hereafter S13; Soubiran et al. 2018). The stability of those stars are good enough for almost all existing/planning spectroscopic surveys. However, the limited number as well as the relatively bright limiting magnitude (V ∼ 9–10 mag, see Section 2) of the stars prevent reliable determinations of RVZPs for most of the existing/planning surveys given that only a very limited number of standard stars are targeted by the surveys.

To solve the problem, we attempt to construct a new set of RV standard stars from the APOGEE data based on the following considerations. First, the high resolution (R ∼ 22,500) and high signal-to-noise ratios (S/Ns) near-infrared (H band; 1.51–1.70 μm) spectra collected by SDSS/APOGEE deliver stellar RV measurements of a precision of ∼100 m s⁻¹ (Holtzman et al. 2015). Second, the primary science targets selected by the survey lie in the magnitude range between 7 ≤ H ≤ 13.8 mag. The bright limit of this range yields sufficient number of stars in common with the existing RV standard stars allowing a robust determination of the RVZP of the APOGEE survey itself, whereas the faint limit ensures that a large number of APOGEE RV standard stars thus built can be targeted by other spectroscopic surveys. Third, the recently released SDSS DR14 (Abolfathi et al. 2017) includes APOGEE observations of over one million spectra of 277 371 unique stars collected by APOGEE-1 (2011 September–2014 July) and APOGEE-2 (2014 July–2016 July) surveys. Over 80% of the 277 371 unique stars have been visited by two times or more. The very high fraction (≥80%) of multi-epoch observations for a large number (over two hundred thousand) of stars with a long time baseline (nearly five years) allows one to select and build a large sample of RV standard stars.

The paper is organized as follows. In Section 2, we describe the reference RV standard stars selected from the existing databases. In Section 3 we describe in detail the construction of a sample of RV standard stars from the APOGEE data. As an application, we use the new set of standard stars to examine the RVZPs of several finished/ongoing large-scale spectroscopic surveys in Section 4. We summarize in Section 5.

In this section, we attempt to build a set of reference RV standard stars selected from the existing databases. These reference RV standard stars will be used to calibrate the RV measurements of the APOGEE survey in the next Section. In doing so, we first briefly introduce the four databases used in the current work as follows.

The first one is the 38 "New" ELODIE-CORAVEL high-precision standard stars (UMQ99), the official set of IAU RV standards. These stars have been observed more than ∼10 times for years and have velocity variations lower than a few 10 m s⁻¹.

The second one, provided by N02, contains 889 late-type stars observed with the HIRES echelle spectrometer on the 10 m Keck I telescope and with the "Hamilton" echelle spectrometer installed on either the 3 m Shane telescope or the 0.6 m Coude Auxilliary Telescope. All those stars are typically visited by 12 times in the period between 1997 and 2001 and 782 of them show velocity variations smaller than 100 m s⁻¹.

The third one, built by C12, is an extension of the N02 database to include more stars (2046 FGKM-type stars in total) observed with the HIRES echelle spectrometer on the 10 m Keck I telescope between 2004 August and 2011 January , as parts of the California Planet Survey (Howard et al. 2010). Among the 2046 FGKM-type stars, 131 are selected as RV standard stars by C12 since they exhibit stable RV for at least 10 years, with a velocity scatter less than 30 m s⁻¹.

Finally, S13 provide a catalog of 1420 potential RV standard stars with data taken from either archives or new observations. Over 90% of them exhibit a stability better than 300 m s⁻¹ over several years. The main goal of the project is to provide a stable RV reference stars to calibrate RV measurements of Gaia-RVS. The project uses data collected with five high-resolution spectrographs: ELODIE (R = 42,000), SOPHIE (R = 75,000), NARVAL (R = 78,000), CORALE (R = 50,000) and HARPS (R = 1,20,000). RV measurements from the different spectrographs are combined, adopting the SOPHIE measurements as the reference frame. The results are compared to values published in other existing catalogs, including N02 and C12. The RV offset between S13 and UMQ99 values reflects the RVZP between SOPHIE and ELODIE frames, and is also reported in S13.

Following S13, we adopt the SOPHIE frame as the RV reference scale when combining the four aforementioned RV databases. The corrections used to correct measurements of the other three databases to S13 (SOPHIE frame) are again taken from S13 and listed in Table 1. The criteria for the selection of the final reference RV standard stars are N_obs ≥ 3; time baseline longer than one year and stability (3σ_RV) better than 100 m s⁻¹. In case a star is available from more than two databases, the preference is that S13 first, UMQ99 second and C12 third. Finally, we obtain a total of 1611 reference RV standard stars, with 29 from UMQ99, 347 from N02, 332 from C12 and 903 from S13 (see Table 1). The spatial distribution of these stars is shown in Figure 1, and is quite homogeneous across the whole sky. The color–magnitude (V against B − V) of the stars is presented in Figure 2. As the plot shows, most of them are FGK-type stars brighter than 10 mag and of colors between 0.5 and 1.0 mag. The sample also includes a few hundred M-type stars fainter than 10 mag and of colors redder than 1.0 mag.

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

3.1. The APOGEE Survey

3.2. RVZP of the APOGEE Instrument

Before selecting RV standard stars from the APOGEE data, we first examine the RVZP of the APOGEE instrument using the reference RV standard stars constructed in Section 2. In doing so, we cross match the reference RV standard stars with the APOGEE DR14⁷ catalog of stellar properties (including RV) deduced from the combined spectra and find 76 stars in common.

Generally, the RVZP of an instrument is affected by a number of factors, including the instrument itself, the spectral resolution and coverage, the wavelength calibration, and the method of deriving RV (e.g., by cross-correlation), and observing conditions and environment (e.g., Gullberg & Lindegren 2002; S13). Provided the environmental conditions are well controlled, the offset between two instruments is typically a constant value or shows some dependencies on the stellar color (or effective temperature), caused by the variations of spectral features used to derive RVs. We therefore examine the offset between the APOGEE and the reference RV measurements as a function of effective temperature (T_eff), as measured from the APOGEE spectra (Figure 3). As the plot shows, the offset shows a clear, nearly linear trend of variations with effective temperature. The offset is about 0.5 km s⁻¹ at T_eff ∼ 6000 K and −0.2 km s⁻¹ at T_eff ∼ 3500 K. To quantitatively describe the trend, a simple linear fit is applied as follows:

$$\begin{eqnarray}&&{\rm{\Delta }}\mathrm{RV}=-1.1069+0.2558\times ({T}_{\mathrm{eff}}/{10}^{3}\,{\rm{K}})\,\mathrm{km}\,{{\rm{s}}}^{-1},\end{eqnarray} \tag{ 1 }$$

where ΔRV is ${\mathrm{RV}}_{\mathrm{APOGEE}}-{\mathrm{RV}}_{\mathrm{REF}}$. This equation will be used to correct the RVs of APOGEE RV standard stars in the next section.

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

The origin of the effective temperature (color) dependence of the offset between the APOGEE RVs and the RVs from the reference standard stars is quite complicated. One probable cause is that their RVs are derived from spectra with different wavelength coverage, namely, the RVs from the APOGEE are derived using near-infrared spectra while the RVs from the reference standard stars are derived using optical spectra.

3.3. Selecting RV Standard Stars from the APOGEE Data

In this section, we attempt to select RV standard stars from the APOGEE data. To do so, we first combine APOGEE RV measurements from the individual spectra obtained in multiple visits as published in SDSS DR14.⁸ Only results from spectra of S/N greater than 50 with RV measurement error smaller than 500 m s⁻¹ are included in the combination. Following S13, we define the following quantities when combining the data:

• 1.  
  Weighted mean RV: $\overline{\mathrm{RV}}=\tfrac{{\sum }_{i=1}^{n}{\mathrm{RV}}_{i}{w}_{i}}{{\sum }_{i=1}^{n}{w}_{i}}$, where w_i is weight as given by the individual RV measurement error: $1/{\epsilon }_{i}^{2}$ and n is the total number of observations;
• 2.  
  Internal error of $\overline{\mathrm{RV}}$: ${I}_{\mathrm{ERV}}=\tfrac{{\sum }_{i=1}^{n}{\epsilon }_{i}{w}_{i}}{{\sum }_{i=1}^{n}{w}_{i}};$
• 3.  
  $\overline{\mathrm{RV}}$ weighted standard deviation:

  $$\begin{eqnarray*}&&{\sigma }_{\mathrm{RV}}^{2}=\tfrac{{\displaystyle \sum }_{i=1}^{n}{w}_{i}}{{\left({\sum }_{i=1}^{n}{w}_{i}\right)}^{2}-{\sum }_{i\,=\,1}^{n}{w}_{i}^{2}}{\sum }_{i}^{n}{w}_{i}{({\mathrm{RV}}_{i}-\overline{\mathrm{RV}})}^{2};\end{eqnarray*}$$
• 4.  
  Uncertainty of $\overline{\mathrm{RV}}$: max(${\sigma }_{\mathrm{RV}}/\sqrt{N}$, ${I}_{\mathrm{RV}}/\sqrt{N}$);
• 5.  
  Time baseline in days: ΔT;
• 6.  
  Mean Modified Julian day (MJD) of the n observations.

A total of about 150,000 stars are selected with at least two high-quality (i.e., S/Ns ≥ 50 and σ_RV ≤ 500 m s⁻¹) observations from the APOGEE data. The distributions of time baseline (ΔT), number of observations (n), and $\overline{\mathrm{RV}}$ weighted standard deviation (σ_RV) of the selected stars are presented in Figure 4. As the plot shows, most stars have ΔT smaller than 200 days (close to the period of one observational season) and n smaller than 5.

Zoom In Zoom Out Reset image size

We further refine the sample of APOGEE RV standard stars with following criteria:

• 1.  
  At least 3 measurements available from high-quality (S/Ns ≥ 50 and σ_RV ≤ 500 m s⁻¹) observations;
• 2.  
  $\overline{\mathrm{RV}}$ weighted standard deviation σ_RV ≤ 200 m s⁻¹ (i.e., stability better than 600 m s⁻¹);
• 3.  
  Time baseline ΔT longer than 200 days.

Here, we note that the RV stability is defined as 3σ_RV. A total of 18080 APOGEE RV standard stars are selected fulfilling the above criteria. The distributions of time baseline (ΔT), number of observations (n) and $\overline{\mathrm{RV}}$ weighted standard deviation (σ_RV) of these RV standard stars are presented in Figure 5. They have a median σ_RV of 80 m s⁻¹ (i.e., a median stability of 240 m s⁻¹), good enough for the calibration of the RVZPs of most of the currently finished/ongoing large-scale Galactic spectroscopic surveys (e.g., SDSS/SEGUE, RAVE, LAMOST). Their spatial distribution is shown in Figure 6. Most of them stars are in the northern sky given that most of the observations are collected at APO. When the observations of APOGEE-2 collected at LCO become more and more, we expect more stars in the south. In Figure 7, we show the color–magnitude (H against J − K_s) of this final sample of RV standard stars. Most stars (≥90%) are redder than 0.5 in J − K_s due to the APOGEE target selection algorithm (Zasowski et al. 2013, 2017). About 10 per cent bluer (J − K_s ≤ 0.5 mag) main-sequence stars are also available, as add-on targets in the APOGEE survey. The cut at faint end (∼12–12.5 mag) in H-band magnitudes of these standard stars is brighter than the limiting magnitude (H ∼ 13.8 mag) of the APOGEE survey simply because of our S/N cut (≥50) when selecting the stars.

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

Finally, we correct the values of $\overline{\mathrm{RV}}$ of these stars for RVZP using Equation (1) and T_eff given by the combined spectra. Properties of the sample stars, including name, H magnitude, J − K_s, T_eff, $\overline{\mathrm{RV}}$, I_ERV, σ_RV, n, uncertainty of $\overline{\mathrm{RV}}$, ΔT and mean MJD, of the final sample of 18080 APOGEE RV standard stars are listed in Table 2.

Table 2.  The Final Sample of APOGEE Radial Velocity Standard Stars

Name	H	J − Ks	Teff	$\overline{\mathrm{RV}}$ a	IERV	σRV	n	Uncertainty	ΔT	Mean MJD
			(K)	(km s−1)	(km s−1)	(km s−1)		(km s−1)	(days)	(−50000)
J00:00:00.68+57:10:23.4	10.13	0.65	4987	−12.335	0.0092	0.1336	3	0.0772	776	6389
J00:00:21.18+61:36:42.1	10.18	1.27	4214	−122.852	0.0059	0.1142	5	0.0511	274	7211
J00:00:24.72+55:18:47.3	10.70	0.98	4298	−128.941	0.0118	0.0940	3	0.0543	776	6389
J00:00:25.61+55:28:51.1	11.20	0.50	5512	−41.319	0.0209	0.1276	3	0.0737	776	6389
J00:00:31.19+70:56:36.5	10.17	1.09	4300	−55.158	0.0047	0.0612	3	0.0353	322	5964
J00:00:34.75+57:23:25.9	11.57	0.47	5657	−17.006	0.0212	0.1584	3	0.0914	776	6389
J00:00:46.32+56:34:05.7	9.83	0.75	4727	−11.147	0.0080	0.0658	3	0.0380	776	6389
J00:00:48.73+56:47:03.1	11.74	0.80	4759	−73.246	0.0158	0.0568	3	0.0328	776	6389
J00:00:51.43+56:15:56.9	10.94	0.69	4912	−60.790	0.0114	0.0273	3	0.0158	776	6389
J00:00:55.91+63:05:05.1	9.92	1.08	4367	−108.744	0.0072	0.1898	5	0.0849	274	7211
⋯	⋯	⋯	⋯	⋯	⋯	⋯	⋯	⋯	⋯	⋯
J05:50:45.34+17:53:46.6	9.55	0.72	4787	35.126	0.0052	0.1103	4	0.0552	281	6432
J05:50:46.37+11:07:47.7	9.48	0.81	3601	−11.462	0.0042	0.1199	9	0.0400	393	6510
J05:50:47.95−03:54:34.5	9.95	0.78	4932	90.759	0.0061	0.1586	3	0.0916	743	6135
J05:50:49.38−04:26:16.8	11.07	0.77	4969	61.452	0.0151	0.0725	3	0.0418	743	6135
J05:50:50.72+16:53:37.3	10.46	0.73	4913	74.542	0.0133	0.0574	3	0.0331	281	6388
J05:50:54.43+26:10:24.0	10.77	1.01	5116	37.375	0.0085	0.0746	5	0.0333	379	6885
J05:50:55.24+17:52:55.3	9.08	1.05	4182	76.459	0.0029	0.1132	4	0.0566	281	6432
J05:50:55.60+17:47:35.5	10.31	0.80	4664	30.372	0.0077	0.0995	4	0.0497	281	6432
J05:51:02.96+52:39:15.3	10.73	0.72	4893	−46.840	0.0111	0.0891	3	0.0514	237	7146
J05:51:04.50+52:24:28.5	9.43	0.72	4887	−19.496	0.0049	0.0290	3	0.0168	237	7146
⋯	⋯	⋯	⋯	⋯	⋯	⋯	⋯	⋯	⋯	⋯

Note.

^aAfter applying the RVZP corrections given by Equation (1).

Only a portion of this table is shown here to demonstrate its form and content. A machine-readable version of the full table is available.

Download table as:  DataTypeset image

As mentioned, RVZPs of large-scale stellar spectroscopic surveys need to be determined and corrected for further studies (e.g., Galactic kinematics and dynamics). In this Section, we attempt to calibrate the RV measurements by determining their RVZPs for three recent stellar spectroscopic surveys: the RAVE, the LAMOST Galactic Spectroscopic Surveys, and the Gaia-RVS survey, using the new APOGEE RV standard star catalog constructed above. The results are presented in Figures 8, and 9, and Table 3, including the median offset ΔRV of RVs measured by the two surveys and those given by the APOGEE RV standard star catalog, the standard deviation of the offset, s.d., the median spectral S/Ns of stars used in the comparison, and the number of stars N used.

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

Table 3.  Comparisons of RVs Yielded by the RAVE and LAMOST Pipelines and Those of the APOGEE RV Standard Stars

Source	ΔRV (km s−1)	s.d. (km s−1)	$\overline{{\rm{S}}/{\rm{N}}}$	N
RAVE (Siebert et al. 2011)	+0.17	1.27	48	352
LAMOST (LSP3; Xiang et al. 2015, 2017)	−2.60	3.86	60	4849
LAMOST (LASP; Luo et al. 2015)	−3.92	3.63	46	3290
Gaia-RVS (Katz et al. 2018)	+0.36	0.59	⋯	10674

Download table as:  ASCIITypeset image

(i) The RAVE survey: The survey (Steinmetz et al. 2006; Kunder et al. 2017) has collected 520 781 medium-resolution (R ∼ 7500) spectra covering the Ca ii triplet regime (8410–8795 Å) of 457,588 unique stars randomly selected from the southern Hemisphere, in nearly ten years (2003–2013) using the multi-object spectrograph 6dF on the 1.2 m UK Schmidt Telescope of the Australian Astronomical Observatory. The determinations of RVs and atmospheric parameters (T_eff, log g, [Fe/H]) for RAVE stars have been described in detail by Siebert et al. (2011) and Kordopatis et al. (2011, 2013), respectively.

To determine the RVZP of RAVE measurements, we cross match the APOGEE RV standard stars with the RAVE DR5 catalog (Kunder et al. 2017) and find 352 common stars of RAVE spectral S/N greater than 10. The stars yield a mean difference ΔRV = 0.17 km s⁻¹ and a standard deviation s.d. = 1.27 km s⁻¹. The standard deviation reveals the RAVE RV measurement errors and is indeed in good agreement with the predicted uncertainty of the RAVE RV in a mean S/N of 48 (Steinmetz et al. 2006). These values are also consistent with the results reported in the previous studies (e.g., Kunder et al. 2017). The RV differences between RAVE stars and APOGEE RV standard stars (ΔRV) as a function of T_eff, log g, [Fe/H], and S/N are also shown in the top panel of Figure 8 and no significant trends are detected for those parameters.

(ii) The LAMOST Galactic Spectroscopic Survey: The survey has hitherto collected nearly eight million (currently the largest spectral database) low-resolution (R ∼ 1800) optical (3700–9000 Å) spectra of S/Ns ≥ 10 during its first five-year Regular Survey (2012–2017) using the LAMOST telescope at Xinglong Observatory. LAMOST is a 4 m, quasi-meridian reflecting Schmidt telescope equipped with 4000 fibers distributed in a field of view of 5° in diameter (Cui et al. 2012). Details about target selections and scientific motivations of the survey can be found in Zhao et al. (2012), Deng et al. (2012), and Liu et al. (2014). Currently, two stellar parameter pipelines, the LAMOST Stellar Parameter at Peking University (LSP3; Xiang et al. 2015, 2017) and the LAMOST Stellar Parameter Pipeline (LASP; Luo et al. 2015), have been developed to derive RVs and basic atmospheric parameters from the collected spectra.

To examine the RVZPs of LAMOST measurements yielded by LSP3 and LASP, we cross match the APOGEE RV standard stars with LAMOST DR3⁹ delivered by LSP3 and LASP, respectively. For the catalog yielded by LSP3, a total of 4849 common stars are found. The mean RV difference is ΔRV = −2.60 km s⁻¹, with a standard deviation s.d. = 3.86 km s⁻¹. The RV differences as a function of T_eff, log g, [Fe/H] and S/N are again shown in the middle panel of Figure 8. and show almost a constant value for those parameters. For the catalog yielded by LASP, a total of 3290 common stars are found. The mean RV difference is ΔRV = −3.92 km s⁻¹, with a standard deviation s.d. = 3.63 km s⁻¹. The relatively large constant offsets between the RVs from the LAMOST pipelines (both LSP3 and LASP) and the RVs from the APOGEE reference standard stars are possibly due to the wavelength calibration of LAMOST spectra. The current RVs from LAMOST are derived with blue spectra while few sky emission lines are available to recalibrate the wavelength calibration in the blue range. The ∼1 km s⁻¹ offset of RVs between LSP3 and LASP is mainly due to the different masks and wavelength ranges used in the RV determinations. As shown in the bottom panel of Figure 8, the RV differences ΔRV show no obvious changes as a function of T_eff, log g, [Fe/H] and S/N.

We note that, given the relatively large random errors, the effective temperature/color dependence systematics are not significant for both RAVE and LAMOST RVs, examined by the APOGEE RV reference stars.

(iii) The Gaia-RVS Survey: In Gaia DR2 (Gaia Collaboration et al. 2018a), median radial velocities for 7,224,631 stars with 3550 ≤ T_eff ≤ 6900 K and G_RVS ≤ 12 mag (i.e., V ≤ 13 mag) are delivered (Katz et al. 2018; Sartoretti et al. 2018), using the spectra collected by the Radial velocity Spectrometer (RVS) instrument on board Gaia. The RVS is a medium resolving power (R ∼ 11500) integral field spectrograph, covering the Ca ii triplet regime (845–872 mm).

Again, to examine the RVZP of Gaia-RVS measurements, we cross match the APOGEE RV standard stars with Gaia DR2 (Gaia Collaboration et al. 2018a) and find 10674 common stars. The mean difference found by these stars is ΔRV = 0.36 km s⁻¹, with a standard deviation s.d. = 0.59 km s⁻¹. The results are in good agreement with the results reported in Katz et al. (2018). More specifically, we show the RV differences as a function of color (G_BP − G_RP), G magnitude, and number of transits (N_obs) in Figure 9. The differences are almost a constant value for color and number of transits but show a clear positive trend of G magnitude. In the bright range (G ≤ 11 mag), the mean difference is close to zero and then increase to ∼500–600 m s⁻¹ at the faint end (G ∼ 14 mag). This trend is also found by Katz et al. (2018). The overall offset reported in Table 3 is mainly due to the large number of faint stars in the common stars.

In addition, given the high-precision of the RV measurements of our APOGEE RV reference stars, we can quantitatively study the precision of Gaia-RVS RV measurements. Doing so, we divide the common stars into different metallicity bins, i.e., [Fe/H] ≥ −0.1, −0.4 ≤ [Fe/H] < −0.1 and [Fe/H] < −0.4. We do not have bins in surface gravity, as most of our RV standard stars are giant stars. For each metallicity bin, we calculate the standard deviation using the common stars as a function of G magnitude bins, for different effective temperature ranges. The results are shown in Figure 10. Generally, the precision of Gaia-RVS RVs decreases with magnitude and effective temperature, increases with metallicity, which are all expected. The typical values of the precision are few hundred m s⁻¹ at the bright range (G ∼ 9–10 mag) and few km s⁻¹ at the faint end (G ∼ 13–14 mag).

Zoom In Zoom Out Reset image size

Using data derived from about one million high-resolution (R ∼ 22500) near-infrared (H band; 1.51–1.70 μm) spectra of around 300,000 unique stars released in the SDSS DR14, we have built a catalog of 18080 RV standard stars. The RVZP of APOGEE instrument is well calibrated to a reference catalog of RV standard stars selected from existing databases. These APOGEE RV standard stars are observed at least three times and have a stability (3σ_RV) better than 600 m s⁻¹ over a time baseline longer than 200 days. The spatial coverage of these stars is currently limited to the northern sky. Once more measurements based on the observations of APOGEE-2 at LCO are released, additional southern RV standard stars should become available in the future. Due to the target selections of APOGEE, most of the APOGEE RV standard stars are red giant stars (J − K_s ≥ 0.5). As add-on sources, about 10 per cent blue main-sequence standard stars are also included. The H-band magnitude range is 7–12.5 mag, and the faint limit is much fainter than the previous RV standard stars.

As an application, we have determined the RVZPs of three large-scale stellar spectroscopic surveys: the RAVE, the LAMOST Spectroscopic Surveys, and the Gaia-RVS survey using this new catalog of APOGEE RV standard stars. By comparing the RVs of APOGEE RV standard stars with these yielded by the RAVE and LAMOST pipelines, a negligible offset (0.17 km s⁻¹) is found for RAVE RV measurements. For LAMOST RVs yielded by LSP3 and LASP pipelines, the offsets are −2.60 km s⁻¹ and −3.92 km s⁻¹, respectively. The offsets possibly come from the wavelength calibration of LAMOST spectra. No obvious variations of velocity measurement differences as a function of T_eff, log g, [Fe/H], and S/N are found in all cases. For Gaia-RVS RVs, the global offset is around 0.36 km s⁻¹, with a standard deviation of 0.59 km s⁻¹. No significant trends of the offsets as a function of color (G_BP − G_RP) and number of transits (N_obs), but a clear positive trend is found for G magnitude. The offset is close to zero at the bright range (G ∼ 9–10 mag) and reaches ∼500–600 m s⁻¹ at the faint end (G ∼ 14 mag). In addition, we quantitatively study the precision of Gaia-RVS RV measurements as a function of metallicity, effective temperature, and magnitude. The results are consistent with the predicted performances of Gaia-RVS.

Currently, our catalog of APOGEE RV standard stars still has two drawbacks: (1) most of the stars are giants, and (2) the spatial coverage is largely in the northern sky. In future, with more observations collected by APOGEE-2 and other high-resolution spectroscopic surveys (e.g., HERMES-GALAH), we can improve the catalog to have more main-sequence stars and an all-sky coverage.

To conclude, the catalog of APOGEE RV standard stars constructed in the current work should be very useful to determine RVZPs of currently ongoing or future large-scale stellar spectroscopic surveys.

This work is supported by the National Key Basic Research Program of China 2014CB845700, the China Postdoctoral Science Foundation 2016M600849, and the National Natural Science Foundation of China U1531244 and 11473001. The LAMOST FELLOWSHIP is supported by Special fund for Advanced Users, budgeted and administrated by Center for Astronomical Mega-Science, Chinese Academy of Sciences (CAMS).

The Guoshoujing Telescope (the Large Sky Area Multi-Object Fiber Spectroscopic Telescope, LAMOST) is a National Major Scientific Project built by the Chinese Academy of Sciences. Funding for the project has been provided by the National Development and Reform Commission. LAMOST is operated and managed by the National Astronomical Observatories, Chinese Academy of Sciences.

This work has made use of data from the European Space Agency (ESA) mission Gaia (https://www.cosmos.esa.int/gaia), processed by the Gaia Data Processing and Analysis Consortium (DPAC; https://www.cosmos.esa.int/web/gaia/dpac/consortium).