The James Webb Space Telescope (JWST) is expected to revolutionize our understanding of the high-redshift universe, and may be able to test the prediction that the first, chemically pristine (Population III) stars are formed with very high characteristic masses. Since isolated Population III stars are likely to be beyond the reach of JWST, small Population III galaxies may offer the best prospects of directly probing the properties of metal-free stars. Here, we present Yggdrasil, a new spectral synthesis code geared toward the first galaxies. Using this model, we explore the JWST imaging detection limits for Population III galaxies and investigate to what extent such objects may be identified based on their JWST colors. We predict that JWST should be able to detect Population III galaxies with stellar population masses as low as similar to 10(5) M(circle dot) at z approximate to 10 in ultra deep exposures. Over limited redshift intervals, it may also be possible to use color criteria to select Population III galaxy candidates for follow-up spectroscopy. The colors of young Population III galaxies dominated by direct starlight can be used to probe the stellar initial mass function (IMF), but this requires almost complete leakage of ionizing photons into the intergalactic medium. The colors of objects dominated by nebular emission show no corresponding IMF sensitivity. We also note that a clean selection of Population III galaxies at z approximate to 7-8 can be achieved by adding two JWST/MIRI filters to the JWST/NIRCam filter sets usually discussed in the context of JWST ultra deep fields.

Population III (pop III) galaxies, made partly or exclusively of metal-free stars, are predicted to exist at high redshifts and may produce very strong Lya emission. A substantial fraction of these Lya photons are likely absorbed in the intergalactic medium at z > 6, but recent simulations suggest that significant Lya emission may be detectable up to z similar to 8.5, i.e. well into the reionization epoch. Here, we argue that high-redshift pop III galaxies with strong Lya emission can be identified in Hubble Space Telescope imaging data because of their unusual colours. We quantify this effect in some of the filters used in Y-band dropout searches for galaxies at z approximate to 8 and find that pop III galaxies with high Lya fluxes may exhibit much bluer J-H colours at z approximate to 8-10 than any normal type of galaxy at these redshifts. This colour signature can arise even if pop III stars account for as little as similar to 10(-3) to 10(-2) of the stellar mass in these galaxies. Some of the anomalously blue objects reported in current Y-band dropout samples do in fact meet the colour criteria for Lya-emitting pop III galaxies.

Small galaxies consisting entirely of Population III (pop III) stars may form at high redshifts, and could constitute one of the best probes of such stars. Here, we explore the prospects of detecting gravitationally lensed pop III galaxies behind the galaxy cluster J0717.5+3745 (J0717) with both the Hubble Space Telescope (HST) and the upcoming James Webb Space Telescope (JWST). By projecting simulated catalogues of pop III galaxies at z similar to 715 through the J0717 magnification maps, we estimate the lensed number counts as a function of flux detection threshold. We find that the ongoing HST survey Cluster Lensing And Supernova survey with Hubble (CLASH), targeting a total of 25 galaxy clusters including J0717, potentially could detect a small number of pop III galaxies if similar to 1 per cent of the baryons in these systems have been converted into pop III stars. Using JWST exposures of J0717, this limit can be pushed to similar to 0.1 per cent of the baryons. Ultradeep JWST observations of unlensed fields are predicted to do somewhat worse, but will be able to probe pop III galaxies with luminosities intermediate between those detectable in HST/CLASH and in JWST observations of J0717. We also explain how current measurements of the galaxy luminosity function at z = 710 can be used to constrain pop III galaxy models with very high star formation efficiencies (similar to 10 per cent of the baryons converted into pop III stars).

The first Population III stars are predicted to form in minihaloes at z approximate to 10-30. The James Webb Space Telescope (JWST), tentatively scheduled for launch in 2018, will probably be able to detect some of the first galaxies, but whether it will also be able to detect the first stars remains more doubtful. Here, we explore the prospects of detecting an isolated Population III star or a small cluster of Population III stars down to z = 2 in either lensed or unlensed fields. Our calculations are based on realistic stellar atmospheres and take into account the potential flux contribution from the surrounding HII region. We find that unlensed Population III stars are beyond the reach of JWST, and that even lensed Population III stars will be extremely difficult to detect. However, the main problem with the latter approach is not necessarily that the lensed stars are too faint, but that their surface number densities are too low. To detect even one 60 M-circle dot Population III star when pointing JWST through the galaxy cluster MACS J0717.5+3745, the lensing cluster with the largest Einstein radius detected so far, the cosmic star formation rate of Population III stars would need to be approximately an order of magnitude higher than predicted by the most optimistic current models.

Population III galaxies are predicted to exist at high redshifts and may be rendered sufficiently bright for detection with current telescopes when gravitationally lensed by a foreground galaxy cluster. Population III galaxies that exhibit strong Lyα emission should furthermore be identifiable from broadband photometry because of their unusual colors. Here, we report on a search for such objects at z > 6 in the imaging data from the Cluster Lensing And Supernova survey with Hubble (CLASH), covering 25 galaxy clusters in 16 filters. Our selection algorithm returns five singly-imaged candidates with Lyα-like color signatures, for which ground-based spectroscopy with current 8-10 m class telescopes should be able to test the predicted strength of the Lyα line. None of these five objects have been included in previous CLASH compilations of high-redshift galaxy candidates. However, when large grids of spectral synthesis models are applied to the study of these objects, we find that only two of these candidates are significantly better fitted by Population III models than by more mundane, low-metallicity stellar populations.

Supernovae are important probes of the properties of stars at high redshifts because they can be detected at early epochs and their masses can be inferred from their light curves. Finding the first cosmic explosions in the universe will only be possible with the James Webb Space Telescope, the Wide-Field Infrared Survey Telescope and the next generation of extremely large telescopes. But strong gravitational lensing by massive clusters, like those in the Cluster Lensing and Supernova Survey with Hubble (CLASH), could reveal such events now by magnifying their flux by factors of 10 or more. We find that CLASH will likely discover at least 2 - 3 core-collapse supernovae at 5 < z < 12 and perhaps as many as ten. Future surveys of cluster lenses similar in scope to CLASH by the James Webb Space Telescope might find hundreds of these events at z < 15 - 17. Besides revealing the masses of early stars, these ancient supernovae could also constrain cosmic star formation rates in the era of first galaxy formation.

While the Lyα emission line has served as an important tool in the study of galaxies at z < 6.5, Lyα emitters (LAE) have proved to be elusive at higher redshifts. If galaxies exhibiting high Lyα equivalent widths exist at z > 6.5, such rare objects may potentially stand out in multiband imaging surveys because of their unusual colors. We have conducted a search for such objects in the Cluster Lensing And Supernova survey with Hubble (CLASH) survey, and here report on the discovery of three gravitationally lensed images of a single LAE candidate behind the Abell 2261 cluster. Model fits to the CLASH broadband photometry suggests very strong Lyα emission (rest-frame Lyα equivalent width EW(Lyα) > 200 Å, prior to any IGM correction) at a redshift of z ≈ 7.8.