From cosmic dust to planet formation : Building new dust models.

The characterisation of cosmic dust properties is key for understanding, among other things, planet formation processes. Astronomical observations provide us information from which it is possible, but not trivial, to deduce physical properties of cosmic dust. For instance, recent observations of 12 young protostars found dust emissivity indices with values β < 1 [Maury et al. 2019, Galametz et al. 2019, Cacciapuoti et al. 2025], which would imply that dust coagulated into grains over 100µm in size [Ysard et al. 2019], much larger than what predicts actual paradigms of planet formation at this stage of stellar evolution. However, relating the grain sizes to their opacity measured in the millimetre bands is not straightforward and rely heavily on the validity of current dust models used as astrophysical analogues in the community. For example, the optical properties of large dust aggregates in cold environments, as observed in millimetre wavelengths were not explored in a systematic way, limiting the astrophysical interpretation that can be done from the measurements, especially for the dense ISM. Our work addresses this blind spot, building new physically-motivated dust models to interpret the dust signatures in protostellar environments.

Our study concentrates on building realistic dust models of evolved, structurally detailed aggregate dust grains. We use the Discrete Dipole approximation (DDA) code ADDA [Yurkin et al. 2011] to compute the optical properties in extinction, absorption and scattering of our grains, based on laboratory-measured material properties from the THEMIS 2 dust model [Ysard et al. 2024]. Our first study [Carpine et al. 2025] highlighted the heavy dependence of dust optical properties on the shape, but also on the composition of dust grains. We stress here the importance of correct representation of potentially large fluffy aggregates in protostars, when current models in the literature include mostly small compact grains, fitted for the diffuse ISM. Radiative transfer simulations we conducted [Carpine et al. 2026, subm.] showed that none of the various existing ISM dust models are able to reproduce the low emissivity indices values observed in some protostars. Using discrete dipole approximation, along with a novel method we developed to speed up drastically the computation of aggregates optical properties [Carpine et al. 2026b, subm.], we are creating the first dense medium dust database of complex aggregates, based on laboratory-measured material properties THEMIS 2 [Ysard et al. 2024], to provide a physically motivated model to precisely interpret protostellar observations.

Building reliable dust models is decisive in the interpretation of observations of the dense ISM, in our understanding of dust evolution towards planet formation. With our models for aggregate dust grain, different optical properties are inferred, challenging fiducial interpretation and opening the path to robust interpretation of the dust properties during the disk formation stage