Cosmic dust has been observed both in emission and extinction everywhere in our Galaxy as well as in many other galaxies. We now know that dust is mainly composed of carbonaceous and silicate particles sized between a few molecules up to a few hundreds nanometers.

But where is dust formed?
Dust grains are classically thought to form in the winds of Asymptotic Giant Branch (AGB) stars. However, nowadays there is increasing observational evidence for notable dust formation in Supernovae (SNe). This is supported by nucleation models (e.g. Marassi et al. 2015) where dust is formed in dense regions during the first phases of SN explosions. However, not all the dust observed in young SNe is released into the interstellar medium. During the explosion of a SN, a violent forward shock invests the circumstellar medium and a subsequent reverse shock heading to the centre of the ejecta is triggered. The passage of the reverse shock heats the gas and leads to the partial destruction of dust grains mainly due to collisions between energetic gas particles and dust grains (sputtering).
A team of researchers at INAF - Osservatorio Astrofisico di Arcetri and INAF - Osservatorio Astronomico di Roma led by Marco Bocchio, and including Simone Bianchi e Raffaella Schneider, recently developed a new code, GRASH Rev, that allows to follow the dynamics of dust grains in the shocked SN ejecta and to compute the time evolution of the mass, composition and size distribution of the grains. Their work is presented in a paper recently published by A&A (Bocchio et al. 2016).

They consider four well studied SNe in the Milky Way and Large Magellanic Cloud: SN 1987a, Cas A, the Crab
Nebula, and N49. These sources have been observed with both Spitzer and Herschel and the multi-wavelength data allow to better assess the mass of warm and cold dust associated with the ejecta. For each SN, they compute the properties of freshly formed dust using a dust formation model (Marassi et al. 2015) and include these grains as input for the GRASH Rev model.
For all the simulated models, they find good agreement with observations. Their study suggests that SN 1987A is too young for the reverse shock to have affected the dust mass. Hence the observed dust mass of 0.7−0.9M_⊙ in this source can be safely considered as indicative of the mass of freshly formed dust in SN ejecta. Conversely, in the other three SNe, the reverse shock has already destroyed between 10 and 40% of the initial dust mass. However, the largest dust mass destruction is predicted to occur between 10³ and 10⁵ yr after the explosions. Since the oldest SN in the sample has an estimated age of 4800 yr, current observations can only provide an upper limit to the effective dust yields. They find that between 1 and 8% of the observed mass will survive.
In the Milky Way, the timescale between the explosion of two SNe has been estimated to be  40 yr. Using the mean effective dust yield that they have obtained leads to a SN dust production rate almost an order of magnitude larger than the observed dust production rate by AGB stars in the LMC.

However, interstellar dust destruction by SNe is observed and modeled to be larger than the dust formation rate by SNe by a factor > 100, therefore requiring dust accretion in the gas phase. The long-standing conundrum is still standing.

Figure 1: Dust mass evolution as a function of time for the four SNe considered in this study (1987A: black, CasA: green, Crab: blue, and N49: red). Solid lines represent GRASH Rev results and dashed lines the results obtained by BS07 model using the same initial conditions. Data points represent the observed dust masses, and the shaded region indicate the time interval when dust processing fades out.