MaNGA Stellar Library (MaStar)

Example spectra from the MaStar library. Image Credit: S. Meneses-Goytia, R. Yan, and the SDSS collaboration.
Example spectra from the MaStar library. Image Credit: S. Meneses-Goytia, R. Yan, and the SDSS collaboration.

As one component of SDSS-IV, we are building a stellar spectral library with a very comprehensive stellar parameter coverage, a large sample size, and high quality calibrations, using the same instrument as used by the MaNGA survey.

A stellar library is a collection of spectra of individual stars. It is a general tool in astronomy and is used by astronomers working on a wide variety of scientific topics, including exoplanets, stellar astrophysics, galactic and extragalactic astronomy, and cosmology. Our library will be especially useful for the purpose of modeling the stellar population in external galaxies, and particularly those observed by MaNGA as they are observed with the same instrument.

The fiber plugplate system in SDSS-IV enables parallel spectroscopic observations between infrared and optical. At the same time when the APOGEE-2 survey is observing with their infrared spectrograph during bright time, we can take optical spectra of other stars in the same field of view with the MaNGA fiber bundles. The large amount of observing time and the wide sky coverage of the APOGEE-2 survey enable us to build a large, comprehensive, and homogeneous stellar spectral library in the optical.

Technical details

  • Bright-time parallel observations
  • Fall 2014 – Spring 2020
  • 17 science targets and 12 calibration standards per 7 deg2 plate
  • Wavelength: 362-1035 nm, resolution R~2000
  • More than 8000 stars in wide areas of the sky
  • Typically 2-3 epochs with a total of 1 hour exposure per epoch
  • Magnitude range between 11.7 and 17.5 in g-band or i-band
  • S/N ranges from 50-150 per Angstrom

MaStar stellar library spectra will be made publicly available in future data releases.

Target Selection

A stellar library needs to be comprehensive in terms of coverage of different stellar types and stellar parameters. We strive to cover as wide a range as possible in each of the four stellar parameters: effective temperature (Teff), surface gravity (log g), metallicity ([Fe/H]), and the alpha-elements-to-iron ratio ([α/Fe]). We achieve this by targeting stars from existing stellar parameter catalogs from a variety of spectroscopic surveys, including APOGEE-1, APOGEE-2, SDSS/SEGUE, and LAMOST. We also estimated stellar parameters for millions of stars using PanSTARRS1 and APASS photometry and we use these to supplement extreme regions of stellar parameter space. This provides us a comprehensive stellar parameter coverage for our library.


  • Renbin Yan (University of Kentucky) (PI)
  • Yanping Chen (NYU Abu Dhabi)
  • Daniel Lazarz
    (University of Kentucky)
  • Claudia Maraston
    (University of Portsmouth)
  • Sofia Meneses-Goytia (University of Portsmouth)
  • Dmitry Bizyaev (NMSU, APO)
  • Guy Stringfellow
    (University of Colorado, Boulder)
  • Jesus Falcon Barroso (IAC)
  • Zheng Zheng (NAOC)
  • Eddie Schlafly (LBNL)
  • Kyle McCarthy (Industry)
  • MaNGA and APOGEE operation teams

What’s New About MaStar?

How is this different from previous stellar spectroscopic surveys, such as SDSS/SEGUE and LAMOST?

Stellar spectral libraries need to have high quality calibrations, including flux calibration, wavelength calibration, and line spread function (LSF) calibration. Although millions of stars have been observed by previous generations of SDSS and by LAMOST, those spectra are taken with single fibers which do not capture all the flux. Due to alignment errors and atmospheric effects, it is very difficult, if not impossible, for single-fiber spectroscopy to achieve a relative flux calibration accuracy to better than 10% for individual stellar spectra. Using fiber bundles, we can achieve much more accurate flux calibration, as the multiple fibers in a bundle allow us to accurately determine the fraction of light covered by each fiber as a function of wavelength.

Follow us on Twitter: @MaStar_library

In addition, we will obtain spectra with much higher signal-to-noise ratio and wider wavelength coverage than SDSS/SEGUE and LAMOST.