Carbon-Enhanced Ultra Metal-Poor Stars

Updated 24 January 2026

Carbon-Enhanced Ultra Metal-Poor stars are defined by [Fe/H] < -4.0 and [C/Fe] ≥ +0.7, serving as fossil records of Population III nucleosynthesis and early star formation.
Their unique abundance patterns, marked by low neutron-capture elements, link to faint core-collapse supernovae and spinstar mechanisms that shaped early chemical evolution.
High-resolution spectroscopic observations enable precise mapping of these stars in Galactic halo and ultrafaint dwarf galaxies, offering vital insights into early chemical enrichment.

Carbon-Enhanced Ultra Metal-Poor (CEMP-UMP) stars are a distinct subclass of unevolved halo stars that reveal the nucleosynthetic processes and chemical inventory of the earliest stellar generations. Defined by both extremely low iron content ( $\mathrm{[Fe/H]} < -4.0$ ) and significant carbon enrichment ( $\mathrm{[C/Fe]} \gtrsim +0.7$ ), these objects dominate the census of the lowest-metallicity Galactic field and ultra-faint dwarf galaxy populations. The majority of CEMP-UMP stars exhibit little to no enhancement of neutron-capture elements, designating them as "CEMP-no" stars. Their detailed abundance patterns serve as direct fossil records of Population III (Pop III) supernova explosions, early gas cooling, and the process of low-mass star formation in the high-redshift Universe.

1. Definitions and Classification Schemes

The abundance ratio for any pair of elements X and Y is defined as: $[\mathrm{X}/\mathrm{Y}] = \log_{10} \left( \frac{N_X}{N_Y} \right)_\star - \log_{10} \left( \frac{N_X}{N_Y} \right)_\odot$ where $N_X$ is the number density of element X. The absolute abundance is given by: $A(\mathrm{X}) = \log_{10} \left( \frac{N_X}{N_H} \right) + 12.00$ The ultra metal-poor (UMP) regime corresponds to $\mathrm{[Fe/H]} < -4.0$ . CEMP stars are typically those with $\mathrm{[C/Fe]} \geq +0.7$ (Aoki et al. 2007).

CEMP subtypes reflect neutron-capture element enrichment patterns (Beers & Christlieb 2005, Goswami et al. 2023):

CEMP-s: $\mathrm{[Ba/Fe]} \geq +1.0$ , $\mathrm{[Ba/Eu]} > +0.5$ (strong s-process, AGB mass transfer signature).
CEMP-r/s: $\mathrm{[Ba/Fe]} \geq +1.0$ , $\mathrm{[Eu/Fe]} \geq +1.0$ , $0.0 \leq \mathrm{[Ba/Eu]} \leq 1.0$ (i-process/neutron-rich).
CEMP-no: $\mathrm{[C/Fe]} \geq +0.7$ , $\mathrm{[Ba/Fe]} < 0.0$ (no neutron-capture enhancement, Pop III or faint SN origin).

At $\mathrm{[Fe/H]} < -4.0$ , CEMP-no stars constitute more than 75% of the known population (Goswami et al., 2023, Hansen et al., 2015, Placco et al., 2015, Hansen et al., 2014, Placco et al., 22 Jan 2026).

2. Observational Properties and Abundance Patterns

CEMP-UMP stars exhibit a range of characteristic abundance patterns:

Star	[Fe/H]	[C/Fe]	A(C)	[Ba/Fe]	[Sr/Fe]	A(Li)	CEMP Subclass
HE 0134–1519	–3.98	+1.00	5.45	<–0.50	–0.30	1.27	CEMP-no
HE 0233–0343	–4.68	+3.48	7.23	<+0.80	+0.32	1.77	CEMP-no
HE 0144–4657	–4.11	+2.20	...	–0.89	–0.30	...	CEMP-no
LAMOST J125346.09	–4.02	+1.59	...	<–0.30	+0.19	1.80	CEMP-no
J173403+644632	–4.30	+3.10	7.23	...	...	...	CEMP-no
EriIV-9808	–3.25	+1.07	5.69	–0.63	–1.13	...	CEMP-no

Carbon: [C/Fe] enhancements range from +0.7 to +3.5, with absolute abundances $A(\mathrm{C})$ either near the "low-C plateau" ( $A(\mathrm{C}) \approx 6.2-6.5$ ) or above, depending on evolutionary state and progenitor channel (Hansen et al., 2015, Aguado et al., 2017, Hansen et al., 2014).
Neutron-capture elements: CEMP-no stars invariably have low or subsolar [Sr/Fe] and [Ba/Fe], the latter typically with upper limits, and in some cases a "floor" at $A(\mathrm{Ba}) \approx -2.0$ (Hansen et al., 2015, Hansen et al., 2014).
Alpha elements and Fe-peak: [Mg/Fe], [Ca/Fe], [Ti/Fe] are found to be moderately enhanced ( $\approx+0.2$ to $+0.4$ ), consonant with core-collapse SN yields at the lowest metallicity (Hansen et al., 2014, Heiger et al., 14 Aug 2025).
Lithium: Several CEMP-no UMP stars have measured $A(\mathrm{Li})\approx1.3-1.8$ , below the Spite plateau ( $A(\mathrm{Li})\approx2.2$ ), indicating partial depletion either through pre-stellar astration or in situ mixing (Hansen et al., 2014, Li et al., 2015).

3. Nucleosynthetic Channels and Formation Scenarios

Multiple enrichment channels are implicated in the origin of CEMP-UMP stars:

Faint Core-Collapse Supernovae: Models with extensive mixing and fallback eject carbon-rich but iron-poor material, naturally yielding high [C/Fe], low [Fe/H], and minimal $n$ -capture enhancement (Placco et al., 2015, Hansen et al., 2014, Placco et al., 22 Jan 2026). Quantitative fits to Population III SN yields suggest progenitor masses of $20-50~M_\odot$ with explosion energies $0.3-0.9\times10^{51}$ erg can match the observed CNO patterns (Placco et al., 2015, Placco et al., 22 Jan 2026).
Fast-Rotating Massive "Spinstars": Rotation-induced mixing in massive stars enables strong CNO and weak $s$ -process production, with [C/Fe] and [N/Fe] significantly elevated. Spinstars can also account for the large star-to-star scatter in $n$ -capture (Sr, Ba) in the UMP regime (Hansen et al., 2014, Hansen et al., 2015, Hansen et al., 2015, Goswami et al., 2023).
AGB Binary Mass Transfer: While the CEMP-s subtype tracks directly to AGB mass transfer and s-process enrichment, this channel is disfavored for nearly all UMP stars, especially those with [Ba/Fe]<0 and lacking binary signatures (Goswami et al., 2023, Salvadori et al., 2015, Placco et al., 22 Jan 2026). Suda et al. (2016) models demonstrate that a continuous sequence of [Ba/C] can arise via AGB wind accretion and dilution, but this mechanism saturates at [Fe/H] $\sim$ –3 for CEMP-no abundances (Suda et al., 2016).
Inhomogeneous Metal Mixing: Hartwig & Yoshida (2019) show that, under inhomogeneous mixing conditions ( $\Delta[\mathrm{C/H}]\gtrsim1$ dex over $\sim$ 10–30 pc), local survivorship bias can favor the rapid collapse of carbon-rich pockets which fragment before their C-normal neighbors, forming CEMP-no stars from Pop III-normal SNe by selective cooling (Hartwig et al., 2018).

4. Environmental Context and Dwarf Galaxies

CEMP-UMP stars are preferentially found in environments conducive to the retention of faint SN or spinstar ejecta:

Ultrafaint Dwarf Galaxies (UFDs): High-resolution spectroscopy of stars such as EriIV-9808 ( $\mathrm{[Fe/H]} = -3.25$ , $\mathrm{[C/Fe]}=+1.07$ , $\mathrm{[Ba/Fe]}=-0.63$ ) demonstrates the near-universal neutron-capture deficiency and high carbonicity at the lowest metallicity, indicating stochastic or bursty chemical evolution possibly with multi-enrichment (Heiger et al., 14 Aug 2025).
Halo Substructures: The discovery of CEMP-no UMPs like HE 0144–4657 ( $\mathrm{[Fe/H]} = -4.11$ , $\mathrm{[C/Fe]} = +2.20$ ) associated dynamically with the Helmi Stream suggests their origin in dissolved accreted dwarf galaxies or proto-UFDs assimilated by massive progenitors before Milky Way assembly (Placco et al., 22 Jan 2026).
Chemical Evolution Modeling: The high CEMP-no fraction below [Fe/H] $<-4$ in UFDs and the halo, and virtual absence in more massive systems like Sculptor below [Fe/H] $<-3$ , are well-explained by cosmological models restricting Pop III initial masses to 10–140 $M_\odot$ and excluding pair-instability SN channels (Salvadori et al., 2015).

5. The Role of Carbon in Early Gas Cooling and Star Formation

Carbon fine-structure cooling is central to the formation of low-mass stars at the lowest metallicities:

Metallicity Thresholds and Cooling: Simulations with detailed chemical networks (Bovino et al. 2014) show that a critical metallicity $Z_\mathrm{crit} \approx 10^{-3}Z_\odot$ and corresponding carbon abundance permit gas to cool to the cosmic microwave background floor via C II and O I lines, even in the presence of significant far-UV backgrounds (Bovino et al., 2014).
Fragmentation Mass Scales: Under carbon-enhanced conditions, the characteristic Jeans mass for protostar formation drops to $M_J\sim1-10~M_\odot$ , two orders of magnitude lower than in metal-free, H $_2$ -cooling-dominated minihalos (Bovino et al., 2014).
IMF and Faint SN Signature: The existence of low-mass ( $M\lesssim0.8~M_\odot$ ) UMP CEMP-no dwarfs at [Fe/H] $\sim-4.5$ implies that pockets enriched by Pop III faint SNe or spinstars reached the carbonicity required for efficient cooling and fragmentation before being swept up by subsequent high-energy SNe (Aguado et al., 2017, Heiger et al., 14 Aug 2025).

6. Implications for Population III Nucleosynthesis and Galactic Archaeology

CEMP-UMP stars are robust fossil probes for the yields and physics of the first stars:

Constraints on Pop III IMF: The observed abundance trends require that the Pop III IMF include significant numbers of $\sim$ 20–50 $M_\odot$ progenitors exploding as low-energy, mixing-and-fallback SNe. Dominance of pair-instability SNe is ruled out by the survival of extremely C-rich, Fe-poor stars at [Fe/H] $<-4$ (Salvadori et al., 2015, Placco et al., 2015).
Chemical Cartography of the Early Milky Way: The distribution of CEMP-no UMPs in both the field and substructure debris provides tracers for the earliest assembly history of the Milky Way and its accreted satellites (Placco et al., 22 Jan 2026, Goswami et al., 2023).
The Two Carbon Plateaus: The existence of high- and low- $A(\mathrm{C})$ plateaus, with UMP CEMP-no stars populating the latter, constrains (i) the mixing and dilution of Pop III yields, and (ii) the timescales for critical gas cooling and the formation of successive stellar generations (Hansen et al., 2015).
Open Questions: Full 3D, non-LTE abundance modeling; the role of local inhomogeneous mixing in the CEMP-no population fraction; evolutionary corrections in surface abundances for giants; and expanding C and N measurements to larger UMP samples all remain active research fronts (Hansen et al., 2014, Hartwig et al., 2018, Hansen et al., 2015).

7. Current Challenges and Prospects

Progress in CEMP-UMP star research is intimately linked to spectroscopic survey expansion and theoretical advances:

Spectroscopic Methodology: Only high-resolution ( $R\gtrsim30\,000$ ) spectra provide sufficient S/N and precision in the central bands (CH $G$ -band, NH feature, Sr II, Ba II) to distinguish reliably among CEMP subtypes at UMP metallicities. Systematics from 3D/NLTE, evolutionary effects on C/N, and uncertain isotopic ratios ( $^{12}$ C/ $^{13}$ C) remain limiting factors (Hansen et al., 2015, Goswami et al., 2023).
Sample Growth: Ongoing surveys (e.g., LAMOST, SDSS/SEGUE, Pristine, future ELT instrumentation) continually add to the UMP/CEMP-no inventory, deepening the metallicity reach and enabling more robust statistical inferences (Li et al., 2015, Aguado et al., 2017, Hansen et al., 2015).
Multi-Element Abundance Modeling: Automated fitting of observed abundance vectors to Pop III SN yield grids (e.g., “starfit” with the Heger & Woosley 2010 grid) has demonstrated that nearly all UMP CEMP-no stars may be explained by a single faint SN event, but refinement of these models is ongoing (Placco et al., 2015, Placco et al., 22 Jan 2026).
Theoretical Simulations: Validation of scenarios such as inhomogeneous metal mixing and feedback-regulated fragmentation of early halos requires high-resolution 3D hydrodynamical simulations with explicit element tracking and radiative feedback—presently an area of aggressive development (Hartwig et al., 2018).

In summary, carbon-enhanced ultra metal-poor stars, and in particular the CEMP-no subclass, encode the chemical outcomes of metal-free and extremely metal-poor stellar explosions. Their unique spectroscopic signatures and spatial associations illuminate the formation of the first low-mass stars and structure of the early Milky Way, and set essential constraints on the population statistics and nucleosynthetic processes of the first stars. Key areas for future work include expanding high-fidelity UMP samples, advancing non-LTE and 3D abundance models, and integrating chemical evidence with dynamical substructure mapping—critical steps in reconstructing the Galactic evolutionary narrative from the earliest times (Hansen et al., 2014, Placco et al., 22 Jan 2026, Goswami et al., 2023, Hartwig et al., 2018, Hansen et al., 2015, Bovino et al., 2014, Placco et al., 2015, Hansen et al., 2015, Salvadori et al., 2015, Hansen et al., 2015, Suda et al., 2016, Aguado et al., 2017, Li et al., 2015, Heiger et al., 14 Aug 2025).