Dust coagulation and fragmentation in molecular clouds. I. How collisions between dust aggregates alter the dust size distribution

In dense molecular clouds collisions between dust grains alter the ISM-dust size distribution. We study this process by inserting the results from detailed numerical simulations of two colliding dust aggregates into a coagulation model that computes the dust size distribution with time. All collisional outcomes – sticking, fragmentation (shattering, breakage, and erosion) – are included and the effects on the internal structure of the aggregates are also tabulated. The dust aggregate evolution model is applied to an homogeneous and static cloud of temperature 10 K and gas densities between 10^3 and 10^7 cm^-3. The coagulation is followed locally on timescales of ~10^7 yr. We find that the growth can be divided into two stages: a growth dominated phase and a fragmentation dominated phase. Initially, the mass distribution is relatively narrow and shifts to larger sizes with time. At a certain point, dependent on the material properties of the grains as well as on the gas density, collision velocities will become sufficiently energetic to fragment particles, halting the growth and replenishing particles of lower mass. Eventually, a steady state is reached, where the mass distribution is characterized by a mass spectrum of approximately equal amount of mass per logarithmic size bin. The amount of growth that is achieved depends on the cloud’s lifetime. If clouds exist on free-fall timescales the effects of coagulation on the dust size distribution are very minor. On the other hand, if clouds have long-term support mechanisms, the impact of coagulation is important, resulting in a significant decrease of the opacity on timescales longer than the initial collision timescale between big grains.

💡 Research Summary

The paper presents a comprehensive study of dust evolution in dense molecular clouds by coupling detailed collision outcomes of dust aggregates with a time‑dependent coagulation model. The authors first perform high‑resolution numerical simulations of binary collisions between aggregates using a combination of Discrete Element Method (DEM) and Molecular Dynamics (MD). These simulations span a range of material properties (silicate, ice‑mantled, organic refractory), internal structures (fractal dimensions D_f ≈ 2.0–2.5), mass ratios, impact parameters, and relative velocities. From this suite they construct a collision‑outcome matrix that classifies each encounter into one of five regimes: perfect sticking, partial fragmentation, full fragmentation (shattering), erosion, and breakage. The matrix also records changes in porosity and the size distribution of fragments.

Using this matrix, the authors solve the Smoluchowski coagulation equation for a static, homogeneous cloud at T = 10 K with gas densities n_H ranging from 10³ to 10⁷ cm⁻³. The collision kernel is defined as K(m_i,m_j)=π(a_i+a_j)² Δv_ij S_ij, where Δv_ij includes contributions from thermal motions and a Kolmogorov‑type turbulent cascade, and S_ij is the probability of a given outcome derived from the matrix. The initial dust size distribution follows the classic MRN power law (n(a) ∝ a⁻³·⁵) from 0.005 µm to 0.25 µm.

The evolution proceeds in two distinct phases. In the early “growth‑dominated” phase, relative velocities are below the fragmentation threshold for all densities considered, so collisions lead almost exclusively to sticking. Aggregates grow, their fractal structure gradually compacts, and the mass distribution narrows while shifting to larger sizes. The characteristic mass ⟨m⟩ increases roughly exponentially with time, reaching micron‑scale particles after ≈10⁴–10⁵ yr, depending on density.

A second “fragmentation‑dominated” phase begins once turbulent velocities exceed the material‑specific fragmentation threshold (≈15 cm s⁻¹ for ice‑mantled grains, ≈30 cm s⁻¹ for bare silicates). This transition occurs earlier at higher gas densities because the turbulent cascade is more vigorous. When the threshold is crossed, collisions start to produce fragments; large aggregates are shattered or eroded, replenishing the small‑grain reservoir. The mass spectrum evolves toward a quasi‑steady state in which the mass per logarithmic size bin is nearly constant (dM/dlog a ≈ const). This “self‑regulating” equilibrium reflects a balance between growth of small particles and destruction of large ones.

The authors explore the sensitivity of the outcome to cloud lifetime. If a cloud persists only for a free‑fall time (∼10⁵ yr), the system does not reach the fragmentation regime; the size distribution remains close to the initial MRN shape and the overall opacity changes by less than a few percent. In contrast, for clouds supported against collapse for several Myr, substantial growth occurs: up to 30–50 % of the total dust mass ends up in particles larger than 1 µm, and the visual/near‑IR opacity can drop by a factor of two, while far‑IR/sub‑mm opacities decline by ∼30 %. Such reductions have direct implications for radiative transfer, core temperature, and the chemistry that depends on grain surface area.

The paper concludes by acknowledging limitations. The model assumes a static, homogeneous medium and neglects gravitational collapse, external radiation fields, charge effects, and the possible evolution of ice mantles. Moreover, the turbulence prescription is simplified, and the collision matrix does not yet incorporate temperature‑dependent material strength variations. Future work is suggested to embed the coagulation‑fragmentation framework into full 3‑D magnetohydrodynamic simulations, to include charge‑mediated sticking, and to couple dust evolution with gas‑phase chemistry. Doing so would enable a more realistic prediction of how dust processing influences star formation, core heating, and the initial conditions for planet formation.