2020MNRAS.499.4666F

The thermal emission of dust is one of the most important tracers of the interstellar medium: multiwavelength photometry in the far-infrared (FIR) and submillimetre (submm) can be fitted with a model, providing estimates of the dust mass. The fit results depend on the assumed value for FIR/submm opacity, which in most models - due to the scarcity, until recently, of experimental measurements - is extrapolated from shorter wavelengths. Lab measurements of dust analogues, however, show that FIR opacities are usually higher than the values used in models and depend on temperature, which suggests that dust mass estimates may be biased. To test the extent of this bias, we create multiwavelength synthetic photometry for dusty galaxies at different temperatures and redshifts, using experimental results for FIR/submm dust opacity and then we fit the synthetic data using standard dust models. We find that the dust masses recovered by typical models are overestimated by a factor of 2-20, depending on how the experimental opacities are treated. If the experimental dust samples are accurate analogues of interstellar dust, therefore, current dust masses are overestimated by up to a factor of 20. The implications for our understanding of dust, both Galactic and at high redshift, are discussed.