Tom Bakx

Co-developed at Chalmers' GARD group, the ALMA Band 2 receiver opens a completely new way of using ALMA. Using the public commissioning data, I produced the first scientific paper with this receiver, comparing its capability to previous bands and paving the path for future upgrades including the Wideband Sensitivity Upgrade and ALMA2040.

A systematic study of the bright [O III] 88µm line in dusty star-forming galaxies at z = 3–4 from the BEARS sample, connecting the well-studied dusty population at cosmic noon to the [O III]-bright galaxies now routinely confirmed in the Epoch of Reionization.

Combining ~200 hours of ALMA and NOEMA observations across ten spectroscopically confirmed z > 8 galaxies, we performed the deepest dust-stacking experiment in the early Universe to date, finding no significant emission down to average dust masses below 10⁵ solar masses. These stringent limits, consistent across bins of redshift, stellar mass and UV slope, point to inefficient dust build-up in the first 600 million years of cosmic history, likely driven by short time-scales for dust production and rapid removal through galactic outflows.

Using new ALMA Band 9 observations to probe close to the peak of the dust spectrum, we confirmed that MACS0416_Y1 at z = 8.3 harbours exceptionally warm dust (~91 K), making it an ultra-luminous infrared galaxy just 600 million years after the Big Bang — the most distant direct dust detection to date. Its high temperature, rather than a large dust reservoir, drives its extreme infrared luminosity, with 93% of its star formation hidden from UV view. The press release reached over 1 billion people and ranked among the 1% most-viewed studies of 2025.

This paper explored a novel and efficient spectral-line survey technique with ALMA, detecting 66 emission and absorption lines across 19 galaxies in six ALMA bands in just six and a half hours. Future studies will benefit strongly from the planned ALMA Wideband Sensitivity Upgrade.

Most distant galaxies are found through either UV or infrared selection, biasing our understanding. Taking a first step towards a "biasless" selection, I present 14 galaxies within the first billion years found through the robust star-formation tracer [C II]. Future SKA+LOFAR analysis will be key to extending such studies to earlier times.

The discovery of a cosmological overdensity surrounding one of the distant dusty galaxies from my PhD thesis. Spectroscopic follow-up grew into a 40+ hour ALMA project.

Observational verification of the statistical lensing method developed in Bakx et al. 2020, extended to imply a total of ~3000 gravitational lenses across the Herschel fields. Sub-mm-selected lenses are selected purely on the background source, so they probe the foreground mass distribution and offer a complementary test of cosmology.

A fun two-page study of the small disagreement between photometric and spectroscopic results from JWST when finding distant galaxies, which can be attributed to the Eddington bias first described back in 1913.

The first synergy between James Webb and ALMA: we find a robust line at z = 12.117, reporting one of the most distant confirmed galaxies known at the time, and carefully evaluate the statistical methods to pave the way for future JWST/ALMA collaborations.

This study enabled the efficient use of ALMA for one of its key goals: measuring cosmological distances. Together with INFRAVIS and Yale-based sonification experts, I am extending it into an online tool and a sound-based experience that also supports neurodivergent collaborators.

The first direct measurement of dust temperature in the Epoch of Reionization. Previous observations instead measured the dust temperature indirectly; this study showed the power of short-wavelength ALMA observations to characterize dust temperature and reveal the total dust-obscured star formation hidden from HST & JWST.

A dedicated redshift hunt with the IRAM 30m telescope that confirmed distances for four dusty starbursts at z = 3.4–4.1, whose CO luminosities and line widths suggest all are gravitationally lensed. The efficient graphical method for identifying robust spectroscopic redshifts introduced here became the seed of the later High-z Sudoku tool.

One of the most distant detections of ionized carbon. The galaxy shows a very high oxygen-to-carbon luminosity ratio, indicative of harsh radiation fields and/or high gas densities. We find unusually warm dust (> 80 K), which became the starting point for many new studies of dust in the distant Universe.

By cross-matching HerBS with optical (SDSS) and near-infrared (VIKING) surveys, we showed that ~82% of these bright sub-mm sources are gravitationally lensed — and measured, for the first time, that the lensed fraction falls with decreasing 500-µm flux density, exactly as galaxy-evolution models predict.

I produced a catalogue of the 209 brightest, most distant galaxies from the 660 deg² H-ATLAS survey. This catalogue underpins extensive large programs to characterize the redshifts of the Herschel legacy and is the basis of my PhD thesis.