Synthetic photometry of OB star clusters with stochastically sampled IMFs: analysis of models and HST observations

We present a pilot library of synthetic NUV, U, B, V, and I photometry of star clusters with stochastically sampled IMFs and ionized gas for initial masses, M_i = 10³, 10⁴, and 10⁵ M_⊙; t = 1, 3, 4, and 8 Myr; Z = 0.014 and Z = 0.002; and log(U_S) = -2 and -3. We compare the library with predictions from deterministic models and observations of isolated low-mass (<10⁴ M_⊙) star clusters with co-spatial compact H II regions. The clusters are located in NGC 7793, one of the nearest galaxies observed as part of the HST LEGUS and Hα-LEGUS surveys. (1) For model magnitudes that only account for the stars: (a) the residual |deterministic mag - median stochastic mag| can be ≥0.5 mag, even for M_i = 10⁵ M_⊙; and (b) the largest spread in stochastic magnitudes occurs when Wolf-Rayet stars are present. (2) For M_i = 10⁵ M_⊙: (a) the median stochastic mag with gas can be >1.0 mag more luminous than the median stochastic magnitude without gas; and (b) nebular emission lines can contribute with $\gt 50{{\ \rm per\ cent}}$ and $\gt 30{{\ \rm per\ cent}}$ to the total emission in the V and I bands, respectively. (3) Age-dating OB-star clusters via deterministic tracks in the U-B versus V-I plane is highly uncertain at Z = 0.014 for M_i ~ 10³ M_⊙ and Z = 0.002 for M_i ~ 10³-10⁵ M_⊙. (4) For low-mass clusters, the V-band extinction derived with stochastic models significantly depends on the value of log(U_S). (5) The youngest clusters tend to have higher extinction. (6) The majority of clusters have multi-peaked age PDFs. (7) Finally, we discuss the importance of characterizing the true variance in the number of stars per mass bin in nature.