Full spectrum fitting with photometry in PPXF: stellar population versus dynamical masses, non-parametric star formation history and metallicity for 3200 LEGA-C galaxies at redshift z ≈ 0.8

I introduce some improvements to the PPXF method, which measures the stellar and gas kinematics, star formation history (SFH) and chemical composition of galaxies. I describe the new optimization algorithm that PPXF uses and the changes I made to fit both spectra and photometry simultaneously. I apply the updated PPXF method to a sample of 3200 galaxies at redshift 0.6 < z < 1 (median z = 0.76, stellar mass $M_\ast \gtrsim 3\times 10^{10}$ M_⊙), using spectroscopy from the LEGA-C survey (DR3) and 28-bands photometry from two different sources. I compare the masses from new JAM dynamical models with the PPXF stellar population M_* and show the latter are more reliable than previous estimates. I use three different stellar population synthesis (SPS) models in PPXF and both photometric sources. I confirm the main trend of the galaxies' global ages and metallicity [M/H] with stellar velocity dispersion σ_* (or central density), but I also find that [M/H] depends on age at fixed σ_*. The SFHs reveal a sharp transition from star formation to quenching for galaxies with $\lg (\sigma _\ast /\mathrm{km}\, s^{-1})\gtrsim 2.3$ ($\sigma _\ast \gtrsim 200$$\mathrm{km}\, s^{-1}$), or average mass density within 1 kpc $\lg (\Sigma _1^{\rm JAM}/\mathrm{\mathrm{M}_{\odot }kpc^{-2}})\gtrsim 9.9$ ($\Sigma _1^{\rm JAM}\gtrsim 7.9\times 10^9\, \mathrm{\mathrm{M}_{\odot }\ kpc^{-2}}$), or with $[M/H]\gtrsim -0.1$, or with Sersic index $\lg n_{\rm Ser}\gtrsim 0.5$ ($n_{\rm Ser}\gtrsim 3.2$). However, the transition is smoother as a function of M_*. These results are consistent for two SPS models and both photometric sources, but they differ significantly from the third SPS model, which demonstrates the importance of comparing model assumptions.