All the chemical elements around us—from the carbon in our cells to the silicon in our computers—were forged by the stars. Yet despite decades of research, many mysteries remain. We still seek to understand how, where, and when the elements were synthesized throughout the history of the Universe.
Stars act as cosmic factories. In their hot cores, they fuse light elements into heavier ones—but only up to iron. Creating heavy elements beyond iron (the trans-Fe elements) requires a different mechanism: the capture of neutrons. These neutron-capture processes occur in various astrophysical environments where large numbers of free neutrons are available.
Different astrophysical sites contribute in different ways. AGB stars—evolved, expanded stars nearing the end of their lives—produce some heavy elements as they slowly shed their outer layers. Massive stars and supernova explosions enrich space with a wide variety of elements formed during their lifetimes and their dramatic deaths. The most violent events in the cosmos, such as neutron-star mergers, can generate some of the heaviest known elements through intense bursts of neutron capture. Other exotic environments may also contribute, and researchers are still working to identify them.
My research aims to clarify which sites produce which elements, and at what time in cosmic history. By combining observations, models, and laboratory data, we strive to piece together the story of the chemical evolution of the Universe—and understand how the stars built the elements that make life possible. I outline below several key aspects and findings of my work.
Neutron capture processes
Elements heavier than iron are primarily created through neutron-capture reactions. Two main types of these reactions—the slow and the rapid neutron-capture processes (known as the s- and r-processes)—are thought to account for most of the trans-iron elements found in the Universe.
Fig. 1 — Main i-process path (red line) in a 1 M⊙, [Fe/H] = –2.3 AGB model, shown at the base of the second thermal pulse when the neutron density reaches its maximum. The chemical abundances are indicated by the red colour scale (mass fraction), and stable nuclei are marked by black squares. For comparison, typical s- and r-process paths are displayed in blue and green, respectively (from Choplin et al. 2023)
Other secondary neutron-capture processes have also been identified. One of them is the intermediate neutron-capture process, or i-process, which operates at neutron densities between those of the s- and r-processes. The existence of the i-process is supported by observations of stars showing chemical overabundances that cannot be explained by the s- or r-process alone. Instead, their peculiar compositions can be reproduced by models involving an intermediate neutron exposure, as predicted by i-process nucleosynthesis. Some pre-solar grains may also carry isotopic signatures consistent with an i-process origin.
Fig. 2 : Abundances in mass fraction during an i-process (red) and r-process (blue). The black squares are stable nuclei.
The i-process can occur when protons are mixed into a convective helium-burning region, triggering a proton-ingestion event (PIE). Such events may arise in several astrophysical sites, including Asymptotic Giant Branch (AGB) stars (see next section).
Another neutron-capture channel is the n-process, which can develop in the helium-burning shell of massive stars during a supernova explosion (see next sections).
Nucleosynthesis in AGB stars : the s- and i-processes
Before entering the planetary-nebula phase and ending their lives as white dwarfs, stars with initial masses between 0.8 and 8 M⊙ evolve through the Asymptotic Giant Branch (AGB) phase. An AGB star consists of a carbon–oxygen (CO) core surrounded by thin helium- and hydrogen-burning shells and a large convective envelope (Fig. 4). Thermal instabilities in the He-burning shell generate recurrent convective thermal pulses (TPs).
During the periods between pulses, and if some extra mixing (e.g., overshoot) occurs below the convective envelope during the third dredge-up, protons can partially mix into the 12C-rich layers left behind by the TP. This leads to the formation of a 13C pocket, where trans-iron elements are produced through the radiative s-process (Fig. 5). Typical conditions for this radiative s-process are temperatures around T ≈ 100 MK, timescales of about 10⁴ years, and neutron densities of Nₙ ≈ 10⁷ cm-3.
A convective s-process may also develop at the base of the convective TP if the temperature exceeds T ≈ 350 MK, which occurs in AGB stars with initial masses ≳ 3 M⊙ (Fig. 5). The s-process elements synthesized in these regions are eventually mixed to the stellar surface by subsequent third dredge-up events.
Fig. 3 : Ring Nebula captured by JWST. The central star is a 0.6 M☉ post-AGB star that lost its envelope (credits : ESA/Webb, NASA, CSA, M. Barlow, N. Cox, R. Wesson)
Fig. 4 : Schematic (not to scale) image of the structure of an asymptotic giant branch star (figure from Lugaro et al. 2023)
Fig. 6 : Schematic view of the internal structure of an AGB star that experienced a PIE. Grey zones show convective zones. The yellow area corresponds to the CO core. The green (blue) dashed lines show where the energy generated by H burning (He burning) is maximal. The red (magenta) shaded areas show where the s-process (i-process) operates. The numbers indicate different important events: (1) start of PIE, (2) split of the TP, (3) merging of pulse with envelope, and (4) weak PIE (figure from Choplin et al. 2025)
Fig. 7 : Kippenhahn diagram highlighting the important aspects during a PIE in a 1 M⊙, [Fe/H] = - 2.5 AGB model, computed with the STAREVOL code. Grey zones are convective regions. The location of the maximal energy generated by H-burning (He-burning) is indicated by the green (blue) dashed line. The colormap represents the neutron density (from ...)
Fig. 5 : Schematic view of the internal structure of a standard AGB star (without a PIE). Grey zones show convective zones. The yellow area corresponds to the CO core. The green (blue) dashed lines represent the location of the H-burning (He-burning) shell. The red shaded areas show where the radiative and convective s-processes operate (figure from Choplin et al. 2025)
When fully developed, the upper boundary of the convective thermal pulse (TP) can reach into the base of the hydrogen-rich envelope, causing a proton-ingestion event (PIE). The PIE mechanism is illustrated in Figs. 6 and 7. During such an event, protons are mixed into the convective TP, captured by the 12C(p,γ)13N reaction, and rapidly converted into 13C through beta decay. The subsequent 13C(α,n)16O reaction, occurring at temperatures of about 250 MK, produces neutron densities as high as Nₙ ≈ 10^15 cm-3. Under these extreme conditions, i-process nucleosynthesis is triggered and may even synthesize actinides (Fig. 8). The processed material is eventually mixed into the stellar envelope and later expelled through stellar winds.
Since 2021, I have carried out several modelling studies aimed at improving our understanding of the i-process in AGB stars. In particular, I have explored the effects of mass and metallicity, convective overshooting, and nuclear-physics uncertainties on the development and efficiency of the i-process.
Fig. 8 : Synthesis of actinides during i-process nucleosynthesis in an AGB star model.
Heavy-Element Nucleosynthesis in Massive Stars
I also investigate how massive stars forge the heavy elements—from strontium to barium, lead, and the proton-rich isotopes—through a combination of neutron-capture and photodisintegration processes. Massive stars host several distinct nucleosynthesis regimes, whose efficiency can be dramatically altered by rotation, mixing, and explosion dynamics.
Fig. 9 : Schematic view of the pre-supernova
''onion-like'' structure of a massive star
During their hydrostatic evolution, massive stars experience the weak s-process, which operates mainly during core helium burning and later in carbon-shell burning. Rotation strongly enhances this process by mixing freshly produced 12C and 16O into the H-burning shell, boosting primary 14N and subsequently 22Ne, the dominant neutron source via 22Ne(α,n)25Mg. As shown in my grid of 10–150 M⊙ models at low metallicity, rotation can increase the production of first-peak s-elements by 2–3 dex, and in some cases extend the s-process flow toward Ba and even Pb at higher rotation rates or reduced 17O(α,γ) efficiencies. These results demonstrate that rotating massive stars may have played a key role in enriching the early Universe in trans-iron material (Choplin et al. 2018).
Fig. 10 : Jet-like explosion in a fast-rotating massive star (40 M⊙). Shown are the mass particles after 2.8 s. The colour shows the temperature of the mass particles. The inner grey circle shows the central remnant. The black circle with a radius of ∼0.7 × 10^10 cm shows the location of the bottom of the helium shell.
A second process, the n-process, occurs explosively when the supernova shock traverses the helium shell. In rotating progenitors, large amounts of unburnt 22Ne survive until collapse; if the explosion is sufficiently energetic—as in jet-driven hypernova-like events (Fig. 10)—temperatures of ~1–1.5 GK efficiently trigger (α,n) reactions. I did an exploratory work showing that this produces a short but intense neutron burst with densities up to 10^19–10^20 cm⁻³, capable of shifting the pre-existing s-process pattern toward heavier elements (Fig. 11). This combined enhanced s + n scenario can reproduce the abundance signatures of certain carbon-enhanced metal-poor stars showing intermediate (“r/s-like”) patterns, thereby identifying a new astrophysical pathway to these puzzling observations (Choplin et al. 2020).
Massive stars can also synthesize the p-nuclei, the rare proton-rich isotopes beyond iron. Using rotating pre-supernova models and hydrodynamic supernova explosions, I found that the p-process yields are highly sensitive to the star’s rotational history. Rotation enhances the prior s-process and therefore increases the inventory of trans-iron seeds on which the p-process acts. In 25 M⊙ models, fast rotation boosts both s- and p-nucleus yields by 3–4 dex, and a reduced 17O(α,γ) rate leads to efficient production of heavy p-nuclides (A ≳ 140). Explosion energy plays a minor role compared to the pre-existing seed distribution. These exploratory results suggest that rotating core-collapse supernovae, especially at sub-solar metallicities, may contribute far more to the Galactic p-nuclei than previously assumed (Choplin et al. 2022).
Fig. 11 : post-jet mass fractions (for a given mass particle). Black squares denote stable isotopes. The light blue line shows a typical r-process flow for comparison purposes (taken from Arnould & Goriely 2020, their Fig. 5). The grey line shows the neutron drip line. Thin dot-ted lines show the location of the magic numbers.
Overall, these these works highlights massive rotating stars as multi-process nucleosynthesis sites, capable of operating the s-, n-, and p-processes in a connected evolutionary sequence. This provides a coherent framework to explain several long-standing observational signatures in metal-poor stars and to reassess the contribution of massive stars to the chemical evolution of the early Galaxy.