🔭 astrophysics

A Monte Carlo method for tracking dust properties during coagulation in protoplanetary disks

This paper presents a new Monte Carlo coagulation algorithm that strictly conserves the global inventory of dust properties while resolving spatial evolution in protoplanetary disks, thereby eliminating the fluctuations found in existing methods and enabling accurate tracking of complex processes like sublimation and condensation.

Original authors: Nerea Gurrutxaga, Vignesh Vaikundaraman, Joanna Drazkowska

Published 2026-04-29

📖 4 min read☕ Coffee break read

CC BY 4.0

Original authors: Nerea Gurrutxaga, Vignesh Vaikundaraman, Joanna Drazkowska

Original paper licensed under CC BY 4.0 (http://creativecommons.org/licenses/by/4.0/). ✨ This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer

Imagine a protoplanetary disk as a giant, swirling cosmic kitchen where dust grains are the ingredients, and planets are the final cakes being baked. For a planet to form, tiny specks of dust must stick together, growing from the size of a grain of sand to pebbles, and eventually to boulders.

The problem is that tracking every single grain of dust in this cosmic kitchen is impossible for a computer; there are simply too many. So, scientists use a "Monte Carlo" method, which is like using a small, representative sample of grains to guess what the whole crowd is doing.

However, the old way of doing this had a flaw. Imagine you are managing a budget for a group project. If you only update the budget for one person when they swap money with another, the total amount of money in the group might accidentally go up or down due to rounding errors. Over time, these tiny errors pile up, and your budget becomes completely wrong. In the context of space dust, this meant that the total amount of specific materials (like water ice or rock) wasn't being conserved accurately, leading to confusing results about how planets form.

The New Solution: A Strict Accountant

The authors of this paper have built a new "accountant" for these cosmic simulations. Here is how their new method works, using some everyday analogies:

1. Grouping the Crowd
Instead of tracking individual dust grains, the computer groups them into "teams." Each team represents millions of identical grains. When two teams collide, the computer simulates the event for the whole group at once, rather than one grain at a time. This saves a massive amount of computing power.

2. The "Sticking" and "Splitting" Game
When two teams collide and stick together, the new team gets bigger. But here is the trick: the computer doesn't just add them up and move on. It carefully tracks exactly how much "stuff" (mass) moves from one team to another.

The Old Way: If Team A (small) hits Team B (big), the old method might update Team B's stats but leave Team A's stats slightly off, causing a "leak" in the total budget of materials.
The New Way: The new method ensures that if Team A loses a bit of mass, Team B gains exactly that amount. No mass is created or destroyed. It's like a strict accountant ensuring that every penny transferred is recorded perfectly on both sides of the ledger.

3. Managing the "Empty Seats"
As dust grains grow, some teams become empty because all their grains have merged into larger ones. The new method has a clever way to handle this: it takes a large, crowded team and splits it in half to fill the empty seat. This keeps the simulation running smoothly without wasting computer memory on empty groups.

4. The "Water Ice" Test
To prove their method works, the authors ran a test involving water and rock. They simulated a scenario where the disk gets hot (like a sudden burst of energy), causing ice to evaporate, and then cools down, causing the ice to condense back onto the rock.

The Result: The old method lost track of exactly how much water was in the system during these changes, creating "noise" or errors in the data. The new method kept the total amount of water and rock perfectly constant, no matter how much the ice melted or froze.

Why This Matters
This new tool allows scientists to simulate the birth of planets with much higher confidence. It ensures that when they say, "There is exactly this much water in this part of the disk," they aren't just guessing due to computer errors. It's a more precise way to watch the cosmic kitchen mix its ingredients, helping us understand how our own solar system—and others like it—came to be.

The paper concludes that this method is a valuable upgrade for anyone trying to model how dust grows into planets, specifically because it stops the "leaking" of material properties that plagued previous simulations.

1. Problem Statement

Dust coagulation is a fundamental process in planet formation, where micrometer-sized grains grow into pebbles and eventually planetesimals. The efficiency of this growth depends heavily on the physical and chemical properties of dust grains (e.g., porosity, composition, charge).

Current Limitations: Existing Monte Carlo (MC) methods used in planet formation (specifically the approach by Zsom & Dullemond 2008) track representative particles to simulate collisional evolution. However, these methods suffer from a critical flaw: they do not strictly conserve the global inventory of dust properties (such as chemical composition or mass of specific components) across the disk.
The Flaw: In the Zsom & Dullemond approach, when two representative particles collide, only one is updated while the other remains unchanged to maintain a constant number of particles. This asymmetry introduces numerical fluctuations in the composition of dust populations, particularly when particles of vastly different sizes and compositions collide. These fluctuations can grow over time, compromising predictions of dust evolution, especially in models involving sublimation, condensation, or multi-component chemistry.
The Gap: There was a lack of a spatially resolvable MC method that simultaneously enforces strict conservation of global dust property budgets.

2. Methodology

The authors present a new coagulation algorithm that modifies the standard MC approach to ensure strict conservation of mass and composition while resolving spatial evolution.

Core Algorithm

The method groups identical physical particles into "representative groups." A representative particle $i$ represents a group of $N_i$ identical particles with mass $m_i$ . The total mass of the group is $M_i = N_i m_i$ .

Collision Grouping: Instead of simulating single collisions between individual representative particles, the algorithm groups one-to-one collisions. If group $\{i\}$ (with $N_i$ particles) collides with group $\{j\}$ (with $N_j$ particles, where $N_i < N_j$ ), every particle in $\{i\}$ is assumed to stick to a unique particle in $\{j\}$ .
Mass and Number Updates:
- The mass of the surviving representative particle $i$ increases: $m'_i = m_i + m_j$ .
- The number of physical particles in group $j$ decreases: $N'_j = N_j - N_i$ .
- The total mass of the system is strictly conserved because mass is transferred from group $j$ to group $i$ without loss or artificial duplication.
Handling Empty Groups (Merging & Filling):
- When a group is emptied ( $N'_j = 0$ ), its memory is reused.
- Splitting: To maintain resolution, the algorithm preferentially splits the group with the largest total mass ( $M_i$ ) into two, rather than duplicating particles (as done in older methods). This allows the simulation to dynamically resolve the high-mass regime.
- Merging: If a group becomes negligible (mass below a threshold $M_{min}$ ), it is merged with the most similar existing group based on a distance metric in property space (e.g., mass difference or multi-component mass difference). The new representative particle's properties are a weighted average of the merged groups.
Physical Collision Grouping: To improve efficiency, collisions between particles with vastly different mass ratios are grouped. A single simulated event represents multiple physical collisions, calculated based on mass ratios and a grouping parameter ( $dm_{max}$ ).

Key Parameters

Merging Parameter ( $X$ ): A user-defined parameter (Eq. 17) that sets the threshold for merging groups. Lower values of $X$ allow for greater variation in group mass, improving resolution for small grains but requiring more computational resources.
Conservation Logic: In multi-component simulations (e.g., water ice and silicates), the algorithm tracks the mass of each component separately. During sticking, the mass of each component is transferred proportionally, ensuring the total mass of each component is conserved globally.

3. Key Contributions

Strict Conservation: The primary contribution is an algorithm that strictly conserves the global budget of dust properties (mass, composition) during coagulation, eliminating the numerical fluctuations inherent in previous MC methods.
Spatial Resolution: Unlike some global conservation methods that lose spatial detail, this algorithm is designed for spatially resolved simulations in protoplanetary disks.
Dynamic Resolution: By splitting the most massive groups rather than duplicating particles, the method dynamically allocates computational resources to the most evolutionarily significant parts of the size distribution.
Multi-Component Tracking: The method is explicitly formulated to handle multi-component dust (e.g., silicates + water ice), allowing for the tracking of sublimation and condensation processes without violating mass conservation.

4. Results and Validation

The authors validated the method through three distinct tests:

Analytical Solutions (Smoluchowski Equation):
- The algorithm was tested against analytical solutions for constant, linear, and product coagulation kernels.
- High Resolution: Using $10^4$ representative particles, the method reproduced analytical mass distributions over 4–5 orders of magnitude with high accuracy.
- Low Resolution: Even with only 200 particles, the method captured the overall evolution, though deviations appeared for linear and product kernels, suggesting a minimum requirement of ~200 particles for reliable results.
2D Protoplanetary Disk Simulation:
- The method was applied to a 2D disk model including dust transport, growth, and fragmentation, and compared against the standard mcdust code (based on Zsom & Dullemond 2008).
- Comparison: Both methods yielded similar global dust size distributions.
- Advantage: The new method showed improved resolution in the small-grain regime. In the standard method, small grains are poorly resolved because each representative particle carries a fixed total mass. In the new method, group masses vary, allowing more representative particles to sample the small-size population.
- Cost: Computational costs were comparable between the two methods for the tested setups.
Conservation of Components (Water & Silicates):
- A zero-dimensional test simulated an outburst event causing water ice sublimation and subsequent recondensation on silicate grains.
- Fluctuations: The standard Zsom & Dullemond method showed significant numerical fluctuations in the mean ice fraction due to asymmetric updates.
- Conservation: The new method maintained a constant mean ice fraction throughout the simulation, demonstrating strict conservation of both water and silicate mass budgets.

5. Significance

This work provides a robust numerical tool for the next generation of planet formation models.

Chemical Evolution: It enables accurate tracking of complex chemical processes (e.g., volatile sublimation/condensation, oxidation states) where mass conservation of specific elements is critical.
Observational Predictions: By improving the resolution of small dust grains, the method enhances the accuracy of synthetic observations (e.g., opacity calculations) which are often dominated by small particles.
Future Applications: The strict conservation of properties makes this method essential for studying the formation of complex planetary building blocks where the chemical history of the dust is as important as its size.

The code and data are publicly available, facilitating adoption by the broader astrophysics community.