Earth experienced a number of impacts during its accretion stage. The last impact, which is thought to have formed the Moon, determined the initial condition of the Earth. Some geochemical studies indicate that Earth's mantle may not have been mixed throughout its history, suggesting that even the energetic Moon-forming impact did not homogenize the Earth mantle. Based on numerical simulations, we find that if the impactor was Mars-sized, the mantle may remain unmixed during the impact, whereas a more energetic impact could have mixed the mantle based on impact simulations (Nakajima & Stevenson, 2015). This indicate the classical Moon-forming model (i.e. the impactor was Mars-sized) is consistent with the evidence from the Earth interior. You can find related movies here.

But of course this is not end of the story -- even if the mantle remain unmixed during the impact, the mantle heterogeneity must be preserved for billions of years. We are currently investigating how the Earth's interior would have evolved over time using a geodynamical model.



Even though it is widely accepted that the Moon formed by a large impact, it is still not clear if this hypothesis can be reconciled with the observed geochemical constraints. For instance, previous work suggests that the Moon would have lost a significant amount of water because the impact would have been so energetic that all the lunar water was lost to space. This model, however, may contradict geochemical studies that suggest that the Moon retains a significant amount of water.

To understand the connection between the impact model and lunar water abundance, we first run impact simulations to determine the disk structure. Our calculations show that the disk temperature near the mid-plane is 4000-7000 K (Nakajima & Stevenson, 2014). We further determine the disk structure and find that water escape from the disk would have been inefficient, and therefore the impact model is consistent with a water-rich Moon (Nakajima & Stevenson, 2018, in rev). We are further investigating how water and volatiles evolve over time. 


Two Martian moons, Phobos and Deimos, are originally thought to be gravitationally captured asteroids. However, recent studies suggest that this model may not explain their current orbits. Alternatively, these moons could have formed by a large impact, as likely was the case for the Earth's Moon. The next question is how we could differentiate these models based on observable quantities. We suggest that water would play a key role -- if the moons formed by impact, these moons could be depleted in water (Nakajima and Canup, 2016 LPSC and 2017 AGU).


The water abundance can be deduced from the future mission called Mars Moons eXploration Mission (MMX), which is schedule to launch in 2024. I am on the mission as a science member, and I am very much looking forward to what we can hear from these moons!



Water plumes were detected by Cassini near the south polar region of Saturn's satellite Enceladus in 2006. These plumes are thought to be feeding Saturn's E ring and also responsible for the strong infrared radiation from Enceladus. The plumes, which are mixture of water vapor, ice and volatiles, are emitted from cracks ("the tiger stripes") on the south polar region. The dynamics of the plumes, such as how these plumes have been maintained, and how long they last, have been actively debated. We perform dynamical simulations that take into account plume-ice interactions, and we find that plumes originating from a subsurface liquid ocean can explain the observed mass flux, heat flow, and thermal signatures (Nakajima & Ingersoll, 2016; Ingersoll & Nakajima, 2016). 



Owing to the recent advancements of space and ground-based telescopes, a few thousand of exoplanets have been detected. These planets are believed to have grown through collisions with other celestial bodies. Just like our Moon, these collisions may sometimes lead to formation of satellites (exomoons). Finding an exomoon is highly challenging, but recently a new study suggests that they may have detected an exomoon, whose confirmation is expected in the near future. 

We propose that such a satellite formation process strongly depends on the mass and composition of the exoplanet. Our numerical simulations indicate that a massive or ice-rich planet may not be susceptible to moon-formation by giant impacts (Nakajima et al., in prep.)



Impacts are highly important processes to determine the conditions of planets. For example, impacts contribute to the core-mantle differentiation as well as removing and adding volatiles to the planets. Thus,we cannot know the planetary history without knowing its impact record. Up until today, a number of studies have been conducted on impacts using theoretical, numerical, and experimental approaches especially for impact craters, but less is known for large impacts where planetary gravity is important.

We perform giant impact numerical simulations to understand impact physics. We are especially  interested in how much mantle melts by impact, because this would affect core formation as well as determining the planetary composition. We use various numerical codes (e.g. SPH, iSALE) to compare their results.