bib BPASS-in-Browser

Simple stellar populations, and how to read them here

A short guide to the models behind the SSP Explorer, and to the knobs on the page.

1 · What an SSP is

Take one star and physics can tell you how it evolves: how bright and how hot it is at every age. A galaxy, though, contains billions of stars of different masses, born together and evolving at wildly different speeds. Population synthesis answers the practical question: if a million solar masses of gas turns into stars today, what does that population look like after 1 Myr? 100 Myr? 10 Gyr? It evolves a grid of stellar models, weights them by how many stars of each mass are born, and sums their light.

The result for an instantaneous burst of star formation is a simple stellar population (SSP): every star has the same age and the same initial composition; only the birth mass varies. More complicated star-formation histories are sums of SSPs. Everything in the SSP Explorer is one SSP, at 51 ages from 1 Myr to 100 Gyr (0.1 dex steps in log age). Formed mass is a slider.

2 · Choosing a population

Initial mass function

The IMF says how many stars of each mass are born. Massive stars are rare and dominate the ultraviolet and the ionizing continuum; low-mass stars are common and lock up most of the mass. BPASS quotes the high-mass slope α of ξ(log10 M) — the number of stars per decade in mass. Salpeter is α = 1.35, a straight line of slope −2.35 on the plot below. A larger α is a steeper drop and fewer massive stars.

Two shapes are loaded here. The broken power-law (the BPASS default, in the spirit of Kroupa 2001) is shallower below 0.5 M☉ and Salpeter-like above. Chabrier (2003) uses a log-normal turnover at low mass and the same high-mass slope, so it has fewer faint M dwarfs but a similar budget of O and B stars. Mmax of 100 or 300 M☉ only changes the rarest, hottest stars. The v2.2.1 grid also includes steeper and flatter high-mass slopes (α = 1.00 and 1.70). In the explorer, the ultraviolet and ionizing panels move more than the optical when you change the IMF.

Number of stars per solar mass, matched at 1 M☉. Power-law pieces are straight on this log–log plot; Chabrier turns over below about a solar mass. The lower panel is the ratio to the BPASS fiducial (broken power-law, α=1.35, Mmax=300 M☉). A dotted line marks equality. The same slope shares a colour; Mmax=300 is dashed.

Metallicity

The z labels are the initial mass fraction in elements heavier than helium. z020 is Z = 0.020 (about solar in the BPASS convention); z002 is a tenth of that; the grid reaches zem5 (Z = 10−5). Metal-poor stars are hotter and more compact, their winds are weaker, and the population produces more ionizing photons — important for the early Universe.

Alpha enhancement (v2.3)

Not all metals are alike. Core-collapse supernovae make mostly alpha elements (O, Ne, Mg, Si, …); Type Ia supernovae add iron later. Old or rapidly formed populations are therefore alpha-enhanced relative to the Sun. The v2.3 grids give SEDs at five mixes, Δ(log α/Fe) from −0.2 to +0.6, at the same total Z — so an alpha-rich spectrum is also iron-poor. Choose a v2.3 population in the explorer and the alpha menu becomes active.

Single stars and binaries

Most massive stars are born with a close companion. Interaction is then the rule: a giant can spill its envelope onto its partner (Roche-lobe overflow), be stripped to a hot helium core, spin up or rejuvenate the companion, or share a common envelope. In the explorer you can compare single-star and binary populations at the same age, IMF and metallicity. Binary populations stay bluer and harder-ionizing for much longer — single-star light fades in the ultraviolet after about 10 Myr, binary light does not. Mass gainers and mergers can look younger than they are. Supernovae and compact-object mergers occur later, and in different numbers, than single-star models predict.

3 · Dust in front of the population

Dust is a screen in front of the SSP, not a mixture of stars and dust. Each wavelength is multiplied by 10−0.4 Aλ. Light taken out of the spectrum is not put back into the infrared. Four curves are offered: Cardelli, Clayton & Mathis (1989) and Fitzpatrick (1999) for the Milky Way, Gordon et al. (2003) for the SMC bar, and Calzetti et al. (2000) for starbursts. Calzetti published an attenuation law; here it is used as a screen, like the others.

Each curve is only defined over a published wavelength window. Outside that window nothing is applied — the spectrum is left as it is, rather than extrapolated. Photometry through a filter that sits outside, or straddles, the window is labelled in the table. The Milky Way and SMC curves are evaluated with the dust_extinction package; Calzetti is the published polynomial.

The same 10 Myr, solar-metallicity, binary SSP (the BPASS fiducial IMF) with E(B−V) = 0.3. RV = 3.1 where the curve allows it; Calzetti and the SMC bar keep their published values (4.05 and 2.74). Each extinguished spectrum is drawn only where that law is defined.

4 · Looking through the intergalactic medium

At redshift z > 0 the explorer stretches wavelengths by (1+z) and dims the plotted flux to the luminosity distance. Neutral hydrogen along the line of sight then absorbs light blueward of Lyα in the observer frame. That is a separate knob from dust: dust acts on the rest-frame spectrum; IGM absorption acts afterwards, on what a distant observer sees.

Two mean sightline averages are offered — Inoue et al. (2014), the default, and Madau (1995), kept for comparison with older work. Neither is a particular sightline, and neither includes a damping wing from a still-neutral IGM at z ≳ 6. Photometry through a filter that reaches below Lyα(1+z) is labelled IGM when the correction is on, and IGM ignored when it is off.

The same fiducial SSP at 10 Myr, placed at z = 3, with no IGM correction and with each of the two models. The dotted line is observed Lyα. The forest and the Lyman continuum sit to the left of that line.

5 · Using the SSP Explorer

  1. Pick a population — release, single or binary, IMF, metallicity, and (for v2.3) the alpha mix.
  2. Move in age with the slider, or press play. Extra panels (star counts, ionizing photons, supernova rates, surviving mass, yields) share the same age. Frames between tabulated ages are interpolated for the eye only; numbers you read out are always at true grid ages.
  3. Mass is a vertical shift: spectra and star counts scale with formed mass. Redshift stretches the SED and switches photometry from absolute to apparent AB magnitudes.
  4. Pin a population to keep it on the plots while you change the live one. Sync ages and sync redshifts lock every overlay to the live age or the live z. Dust and mass stay as they were when you pinned. Summing visible spectra adds the live SSP to every pin at the ages shown — a mix of spectra, not a star-formation history.
  5. Photometry is for the live SSP. Tick a filter to overlay its passband. The Correction applied column uses short pills: Dust, IGM, or None, with a warning when a requested correction could not be fully applied in that band.
  6. Export the SED plot as a PNG, or the spectrum as CSV or JSON at full native resolution. For tables with units and a query header, use Fetch data.

What is and is not model output

Animation frames between BPASS grid ages are linear interpolations for visual smoothness. Photometry, downloads and the numbers in the tables are evaluated at tabulated ages only.

6 · Units

7 · Further reading

Open the SSP Explorer