Space School UK (SSUK) is a series of summer residential programmes for secondary school aged students, held at the University of Leicester over 3 weeks each year. Each programme involves space-related activities run by a team of mentors - currently including university students, graduates, teachers and young professionals associated with the space sector - all of whom attended SSUK as students themselves. It includes the 6-day Space School UK and the 8-day Senior Space School UK (collectively SSUK) which are for 13-15 and 16-18 year olds respectively. This paper seeks to evaluate and present the benefits of SSUK to individuals who participate in the programme, organisations involved in the running of SSUK, and to highlight and promote these benefits to the wider UK and global space community. We also address which facets of SSUK make for such an engaging and encouraging experience for the students, that are missing from students' traditional education. We seek to show how SSUK acts as an excellent example of how to bridge the gap between secondary and tertiary space education. Through an analysis of our alumni survey results, we show that attending SSUK has a significant impact on career choices and prepares students for Higher Education, regardless of background. Some groups, such as women, and those from non-selective schools, reported a higher impact in some of these areas than others. Metrics such as skills learned, goals achieved, alongside knowledge of careers and Higher Education possibilities are discussed for various demographics.

Abridged: Getting a better understanding of the evolution and nucleosynthetic yields of the most metal-poor stars (appr. Z<=10^-5) is critical because they are part of the big picture of the history of the primitive Universe. Yet many of the remaining unknowns of stellar evolution lie in the birth, life, and death of these objects. We review stellar evolution of intermediate-mass (IMS) Z<=10-5 models existing in the literature, with a focus on the problem of their final fates. The depth and efficiency of mixing episodes are critical to determine the mass limits for the formation of electron-capture supernovae, but our knowledge of these phenomena is not complete because they are strongly affected by the choice of input physics. We also consider the alternative SNI1/2 channel to form SNe out of the most metal-poor IMS. In this case, it is critical to understand the thermally-pulsing AGB evolution until the late stages. Efficient second dredge-up and, later, third dredge-up episodes could be able to pollute stellar envelopes enough for the stars to undergo thermal pulses in a way very similar to that of higher initial Z objects. Inefficient 2nd and/or 3rd dredge-up may leave an almost pristine envelope, unable to sustain strong stellar winds. This may allow the H-exhausted core to grow to M_Ch before the envelope is lost, and thus let the star explode as a SNI1/2. After reviewing the information available on these two possible channels for the formation of SNe, we discuss existing nucleosynthetic yields of stars of metallicity Z<=10^-5, and present an example of nucleosynthetic calculations for a thermally-pulsing Super-AGB star of Z=10^-5. We compare theoretical predictions with observations of the lowest [Fe/H] objects detected. The review closes by discussing current open questions as well as possible fruitful avenues for future research.

The origin of the stellar initial mass function (IMF) is a fundamental issue in the theory of star formation. It is generally fit with a composite power law. Some clues on the progenitors can be found in dense starless cores that have a core mass function (CMF) with a similar shape. In the low-mass end, these mass functions increase with mass, albeit the sample may be somewhat incomplete; in the high-mass end, the mass functions decrease with mass. There is an offset in the turn-over mass between the two mass distributions. The stellar mass for the IMF peak is lower than the corresponding core mass for the CMF peak in the Pipe Nebula by about a factor of three. Smaller offsets are found between the IMF and the CMFs in other nebulae. We suggest that the offset is likely induced during a starburst episode of global star formation which is triggered by the formation of a few O/B stars in the multi-phase media, which naturally emerged through the onset of thermal instability in the cloud-core formation process. We consider the scenario that the ignition of a few massive stars photoionizes the warm medium between the cores, increases the external pressure, reduces their Bonnor?Ebert mass, and triggers the collapse of some previously stable cores. We quantitatively reproduce the IMF in the low-mass end with the assumption of additional rotational fragmentation.

We explore the final fates of massive intermediate-mass stars by computing detailed stellar models from the zero age main sequence until near the end of the thermally pulsing phase. These super-AGB and massive AGB star models are in the mass range between 5.0 and 10.0 Msun for metallicities spanning the range Z=0.02-0.0001. We probe the mass limits M_up, M_n and M_mass, the minimum masses for the onset of carbon burning, the formation of a neutron star, and the iron core-collapse supernovae respectively, to constrain the white dwarf/electron-capture supernova boundary. We provide a theoretical initial to final mass relation for the massive and ultra-massive white dwarfs and specify the mass range for the occurrence of hybrid CO(Ne) white dwarfs. We predict electron-capture supernova (EC-SN) rates for lower metallicities which are significantly lower than existing values from parametric studies in the literature. We conclude the EC-SN channel (for single stars and with the critical assumption being the choice of mass-loss rate) is very narrow in initial mass, at most approximately 0.2 Msun. This implies that between ~ 2-5 per cent of all gravitational collapse supernova are EC-SNe in the metallicity range Z=0.02 to 0.0001. With our choice for mass-loss prescription and computed core growth rates we find, within our metallicity range, that CO cores cannot grow sufficiently massive to undergo a Type 1.5 SN explosion.

We use our new population synthesis code BONNFIRES to test how surface abundances predicted by rotating stellar models depend on the numerical treatment of rotational mixing, such as spatial resolution, temporal resolution and computation of mean molecular weight gradients. We find that even with identical numerical prescriptions for calculating the rotational mixing coefficients in the diffusion equation, different timesteps lead to a deviation of the coefficients and hence surface abundances. We find the surface abundances vary by 10-100% between the model sequences with short timestep of 0.001Myr to model sequences with longer timesteps. Model sequences with stronger surface nitrogen enrichment also have longer main-sequence lifetimes because more hydrogen is mixed to the burning cores. The deviations in main-sequence lifetimes can be as large as 20%. Mathematically speaking, no numerical scheme can give a perfect solution unless infinitesimally small timesteps are used. However, we find that the surface abundances eventually converge within 10% between modelling sequences with sufficiently small timesteps below 0.1Myr. The efficiency of rotational mixing depends on the implemented numerical scheme and critically on the computation of the mean molecular weight gradient. A smoothing function for the mean molecular weight gradient results in stronger rotational mixing. If the discretization scheme or the computational recipe for calculating the mean molecular weight gradient is altered, re-calibration of mixing parameters may be required to fit observations. If we are to properly understand the fundamental physics of rotation in stars, it is crucial that we minimize the uncertainty introduced into stellar evolution models when numerically approximating rotational mixing processes.

Powerful telescopes equipped with multi-fibre or integral field spectrographs combined with detailed models of stellar atmospheres and automated fitting techniques allow for the analysis of large number of stars. These datasets contain a wealth of information that require new analysis techniques to bridge the gap between observations and stellar evolution models. To that end, we develop BONNSAI (BONN Stellar Astrophysics Interface), a Bayesian statistical method, that is capable of comparing all available observables simultaneously to stellar models while taking observed uncertainties and prior knowledge such as initial mass functions and distributions of stellar rotational velocities into account. BONNSAI can be used to (1) determine probability distributions of fundamental stellar parameters such as initial masses and stellar ages from complex datasets, (2) predict stellar parameters that were not yet observationally determined and (3) test stellar models to further advance our understanding of stellar evolution. An important aspect of BONNSAI is that it singles out stars that cannot be reproduced by stellar models through $\chi^{2}$ hypothesis tests and posterior predictive checks. BONNSAI can be used with any set of stellar models and currently supports massive main-sequence single star models of Milky Way and Large and Small Magellanic Cloud composition. We apply our new method to mock stars to demonstrate its functionality and capabilities. In a first application, we use BONNSAI to test the stellar models of Brott et al. (2011a) by comparing the stellar ages inferred for the primary and secondary stars of eclipsing Milky Way binaries. Ages are determined from dynamical masses and radii that are known to better than 3%. We find that the stellar models reproduce the Milky Way binaries well. BONNSAI is available through a web-interface at http://www.astro.uni-bonn.de/stars/bonnsai.

The very bright red star HV2112 in the Small Magellanic Cloud could be a massive Thorne-Zytkow Object, a supergiant-like star with a degenerate neutron core. With its luminosity of over $10^5\,\rm L_\odot$, it could also be a super asymptotic giant branch star, a star with an oxygen/neon core supported by electron degeneracy and undergoing thermal pulses with third dredge up. Both TZOs and SAGB stars are expected to be rare. Abundances of heavy elements in HV2112's atmosphere, as observed to date, do not allow us to distinguish between the two possibilities based on the latest models. Molybdenum and rubidium can be enhanced by both the irp-process in a TZO or by the s-process in SAGB stars. Lithium can be generated by hot bottom burning at the base of the convective envelope in either. HV2112's enhanced calcium could thus be the key determinant. A SAGB star is not able to synthesise its own calcium but it may be possible to produce this in the final stages of the process that forms a TZO, when the degenerate electron core of a giant star is tidally disrupted by a neutron star. Hence our calculations indicate that HV2112 is most likely a genuine TZO.

We present a new grid of stellar models and nucleosynthetic yields for super-AGB stars with metallicities Z=0.001 and 0.0001, applicable for use within galactic chemical evolution models. Contrary to more metal rich stars where hot bottom burning is the main driver of the surface composition, in these lower metallicity models the effect of third dredge-up and corrosive second dredge-up also have a strong impact on the yields. These metal-poor and very metal-poor super-AGB stars create large amounts of He4, C13 and N14, as well as the heavy magnesium isotopes Mg25 and Mg26. There is a transition in yield trends at metallicity Z approximately 0.001, below which we find positive yields of C12, O16, N15, Al27 and Si28, which is not the case for higher metallicities. We explore the large uncertainties derived from wind prescriptions in super-AGB stars, finding approximately 2 orders of magnitude difference in yields of Ne22, Na23, Mg24,25,26, Al27 and our s-process proxy isotope g. We find inclusion of variable composition low temperature molecular opacities is only critical for super-AGB stars of metallicities below Z approximately 0.001. We analyze our results, and those in the literature, to address the question: Are super-AGB stars the polluters responsible for extreme population in the globular cluster NGC 2808? Our results, as well as those from previous studies, seem unable to satisfactorily match the extreme population in this globular cluster.

BONNFIRES, a new generation of population synthesis code, can calculate nuclear reaction, various mixing processes and binary interaction in a timely fashion. We use this new population synthesis code to study the interplay between binary mass transfer and rotation. We aim to compare theoretical models with observations, in particular the surface nitrogen abundance and rotational velocity. Preliminary results show binary interactions may explain the formation of nitrogen-rich slow rotators and nitrogen-poor fast rotators, but more work needs to be done to estimate whether the observed frequencies of those stars can be matched.

We have computed detailed evolution and nucleosynthesis models for super and massive AGB stars over the mass range 6.5-9.0 Msun in divisions of 0.5 Msun with metallicities Z=0.02, 0.008 and 0.004. These calculations, in which we find third dredge-up and hot bottom burning, fill the gap between existing low and intermediate-mass AGB star models and high mass star models that become supernovae. For the considered metallicities, the composition of the yields is largely dominated by the thermodynamic conditions at the base of the convective envelope rather than by the pollution arising from third dredge up. We investigate the effects of various uncertainties, related to the mass-loss rate, mixing length parameter, and the treatment of evolution after the envelope instability that develops near the end of the (Super)AGB phase. Varying these parameters alter the yields mainly because of their impact on the amount of third dredge up enrichment, and to a lesser extent on the hot bottom burning conditions. Our models produce significant amounts of He4, Li7 (depending on the mass-loss formulation) C13, N14, O17, Na23, Mg25, as well the radioactive isotope Al26 in agreement with previous investigation. In addition our results show enrichment of Ne22, Mg26 and Fe60, as well as a substantial increase in our proxy neutron capture species representing all species heavier than iron. These stars may provide important contributions to the Galaxy's inventory of the heavier Mg isotopes, N14, Li7 and Al27.

The J-type carbon (J)-stars constitute 10-15% of the observed carbon stars in both our Galaxy and the Large Magellanic Cloud (LMC). They are characterized by strong 13C absorption bands with low 12C/13C ratios along with other chemical signatures peculiar for typical carbon stars, e.g. a lack of s-process enhancement. Most of the J-stars are dimmer than the N-type carbon stars some of which, by hot-bottom burning, make 13C only in a narrow range of masses. We investigate a binary-star formation channel for J-stars involving re-accretion of carbon-rich nova ejecta on main-sequence companions to low-mass carbon-oxygen white-dwarfs. The subsequent evolution of the companion stars in such systems is studied with a rapid binary evolutionary code to predict chemical signatures of nova pollution in systems which merge into giant single stars. A detailed population synthesis study is performed to estimate the number of these mergers and compare their properties with observed J-stars. Our results predict that such nova polluted mergers evolve with low luminosities as well as low 12C/13C ratios like the majority of observed J-stars (e.g. in the LMC) but cannot account for the observed fraction of J-stars in existing surveys of carbon stars.

Our main goals are to get a deeper insight into the evolution and final fates of intermediate-mass, extremely metal-poor (EMP) stars. We also aim to investigate their C, N, and O yields. Using the Monash University Stellar Evolution code we computed and analysed the evolution of stars of metallicity Z = 10^-5 and masses between 4 and 9 M_sun, from their main sequence until the late thermally pulsing (super) asymptotic giant branch, TP-(S)AGB phase. Our model stars experience a strong C, N, and O envelope enrichment either due to the second dredge-up, the dredge-out phenomenon, or the third dredge-up early during the TP-(S)AGB phase. Their late evolution is therefore similar to that of higher metallicity objects. When using a standard prescription for the mass loss rates during the TP-(S)AGB phase, the computed stars lose most of their envelopes before their cores reach the Chandrasekhar mass, so our standard models do not predict the occurrence of SNI1/2 for Z = 10^-5 stars. However, we find that the reduction of only one order of magnitude in the mass-loss rates, which are particularly uncertain at this metallicity, would prevent the complete ejection of the envelope, allowing the stars to either explode as an SNI1/2 or become an electron-capture SN. Our calculations stop due to an instability near the base of the convective envelope that hampers further convergence and leaves remnant envelope masses between 0.25 M_sun for our 4 M_sun model and 1.5 M_sun for our 9 M_sun model. We present two sets of C, N, and O yields derived from our full calculations and computed under two different assumptions, namely, that the instability causes a practically instant loss of the remnant envelope or that the stars recover and proceed with further thermal pulses. Our results have implications for the early chemical evolution of the Universe.

Certain carbon-enhanced metal-poor stars likely obtained their composition via pollution from some of the earliest generations of asymptotic giant branch stars and as such provide important clues to early Universe nucleosynthesis. Recently, Kinman et al. discovered that the highly carbon- and barium-enriched metal-poor star SDSS J1707+58 is in fact an RR Lyrae pulsator. This gives us an object in a definite evolutionary state where the effects of dilution of material during the Main Sequence are minimised owing to the object having passed through first dredge-up. We perform detailed stellar modelling of putative progenitor systems in which we accreted material from asymptotic giant branch stars in the mass range 1-2 solar masses. We investigate how the surface abundances are affected by the inclusion of mechanisms like thermohaline mixing and gravitational settling. While we are able to find a reasonable fit to the carbon and sodium abundances of SDSS J1707+58, suggesting accretion of around 0.1 solar masses from a 2 solar mass companion, the strontium and barium abundances remain problematic and this object may have experienced something other than a standard $s$ process. We have more success in fitting the abundances of the mildly carbon-enriched, metal-poor RR Lyrae pulsator TY Gru (CS 22881-071), which we suggest received 0.1 solar masses of material from a companion of around 1 solar mass.

We present numerical simulations of internal gravity waves (IGW) in a star with a convective core and extended radiative envelope. We report on amplitudes, spectra, dissipation and consequent angular momentum transport by such waves. We find that these waves are generated efficiently and transport angular momentum on short timescales over large distances. We show that, as in the Earth's atmosphere, IGW drive equatorial flows which change magnitude and direction on short timescales. These results have profound consequences for the observational inferences of massive stars, as well as their long term angular momentum evolution. We suggest IGW angular momentum transport may explain many observational mysteries, such as: the misalignment of hot Jupiters around hot stars, the Be class of stars, Ni enrichment anomalies in massive stars and the non-synchronous orbits of interacting binaries.

We propose that the observed misalignment between extra-solar planets and their hot host stars can be explained by angular momentum transport within the host star. Observations have shown that this misalignment is preferentially around hot stars, which have convective cores and extended radiative envelopes. This situation is amenable to substantial angular momentum transport by internal gravity waves (IGW) generated at the convective-radiative interface. Here we present numerical simulations of this process and show that IGW can modulate the surface rotation of the star. With these two- dimensional simulations we show that IGW could explain the retrograde orbits observed in systems such as HAT-P-6 and HAT-P-7, however, extension to high obliquity objects will await future three- dimensional simulations. We note that these results also imply that individual massive stars should show temporal variations in their v sini measurements.

The literature is rich in analysis and results related to thermally pulsing-asymptotic giant branch (TP-AGB) stars, but the problem of the instabilities that arise and cause the divergence of models during the late stages of their evolution is rarely addressed. We investigate the physical conditions, causes and consequences of the interruption in the calculations of massive AGB stars in the late thermally-pulsing AGB phase. We have thoroughly analysed the physical structure of a solar metallicity 8.5 solar mass star and described the physical conditions at the base of the convective envelope (BCE) just prior to divergence. We find that the local opacity maximum caused by M-shell electrons of Fe-group elements lead to the accumulation of an energy excess, to the departure of thermal equilibrium conditions at the base of the convective envelope and, eventually, to the divergence of the computed models. For the 8.5 solar mass case we present in this work the divergence occurs when the envelope mass is about 2 solar mass. The remaining envelope masses range between somewhat less than 1 and more than 2 solar mass for stars with initial masses between 7 and 10 solar mass and, therefore, our results are relevant for the evolution and yields of super-AGB stars. If the envelope is ejected as a consequence of the instability we are considering, the occurrence of electron-capture supernovae would be avoided at solar metallicity.

Models of rapidly rotating massive stars at low metallicities show significantly different evolution and higher metal yields compared to non-rotating stars. We estimate the spin-down time-scale of rapid rotating non-convective stars supporting an alpha-Omega dynamo. The magnetic dynamo gives rise to mass loss in a magnetically controlled stellar wind and hence stellar spin down owing to loss of angular momentum. The dynamo is maintained by strong horizontal rotation-driven turbulence which dominates over the Parker instability. We calculate the spin-down time-scale and find that it could be relatively short, a small fraction of the main-sequence lifetime. The spin-down time-scale decreases dramatically for higher surface rotations suggesting that rapid rotators may only exhibit such high surface velocities for a short time, only a small fraction of their main-sequence lifetime.

V605 Aquilae is today widely assumed to have been the result of a final helium shell flash occurring on a single post-asymptotic giant branch star. The fact that the outbursting star is in the middle of an old planetary nebula and that the ejecta associated with the outburst is hydrogen deficient supports this diagnosis. However, the material ejected during that outburst is also extremely neon rich, suggesting that it derives from an oxygen-neon-magnesium star, as is the case in the so-called neon novae. We have therefore attempted to construct a scenario that explains all the observations of the nebula and its central star, including the ejecta abundances. We find two scenarios that have the potential to explain the observations, although neither is a perfect match. The first scenario invokes the merger of a main sequence star and a massive oxygen-neon-magnesium white dwarf. The second invokes an oxygen-neon-magnesium classical nova that takes place shortly after a final helium shell flash. The main drawback of the first scenario is the inability to determine whether the ejecta would have the observed composition and whether a merger could result in the observed hydrogen-deficient stellar abundances observed in the star today. The second scenario is based on better understood physics, but, through a population synthesis technique, we determine that its frequency of occurrence should be very low and possibly lower than what is implied by the number of observed systems. While we could not envisage a scenario that naturally explains this object, this is the second final flash star which, upon closer scrutiny, is found to have hydrogen-deficient ejecta with abnormally high neon abundances. These findings are in stark contrast with the predictions of the final helium shell flash and beg for an alternative explanation.

The stellar origin of gamma-ray bursts can be explained by the rapid release of energy in a highly collimated, extremely relativistic jet. This in turn appears to require a rapidly spinning highly magnetised stellar core that collapses into a magnetic neutron star or a black hole within a relatively massive envelope. They appear to be associated with type Ib/c supernovae but, with a birthrate of around 10^{-6}-10^{-5} per year per galaxy, they are considerably rarer than such supernovae in general. To satisfy all these requirements we hypothesize a binary star model that ends with the merging of an oxygen neon white dwarf with the carbon-oxygen core of a naked helium star during a common envelope phase of evolution. The rapid spin and high magnetic field are natural consequences of such a merging. The evolution that leads to these progenitors is convoluted and so naturally occurs only very rarely. To test the hypothesis we evolve a population of progenitors and find that the rate is as required. At low metallicity we calculate that a similar fraction of stars evolve to this point and so would expect the gamma-ray burst rate to correlate with the star formation rate in any galaxy. This too is consistent with observations. These progenitors, being of intermediate mass, differ radically from the usually postulated high-mass stars. Thus we can reconcile observations that the bursts occur close to but not within massive star associations.

We propose here a robust scheme to infer the physical parameters of compact stars from their f-mode gravitational wave signals. We first show that the frequency and the damping rate of f-mode oscillation of compact stars can be expressed in terms of universal functions of stellar mass and moment of inertia. By employing the universality in the f-mode one can then infer accurate values of the mass, the moment of inertia and the radius of a compact star. In addition, we demonstrate that our new scheme works well for both realistic neutron stars and quark stars, and hence provides a unifying way to infer the physical parameters of compact stars.

We investigate the behaviour of asymptotic giant branch (AGB) stars between metallicities Z = 10-4 and Z = 10-8 . We determine which stars undergo an episode of flash-driven mixing, where protons are ingested into the intershell convection zone, as they enter the thermally pulsing AGB phase and which undergo third dredge-up. We find that flash-driven mixing does not occur above a metallicity of Z = 10-5 for any mass of star and that stars above 2 M do not experience this phenomenon at any metallicity. We find carbon ingestion (CI), the mixing of carbon into the tail of hydrogen burning region, occurs in the mass range 2 M to around 4 M . We suggest that CI may be a weak version of the flash-driven mechanism. We also investigate the effects of convective overshooting on the behaviour of these objects. Our models struggle to explain the frequency of CEMP stars that have both significant carbon and nitrogen enhancement. Carbon can be enhanced through flash-driven mixing, CI or just third dredge up. Nitrogen can be enhanced through hot bottom burning and the occurrence of hot dredge-up also converts carbon into nitrogen. The C/N ratio may be a good indicator of the mass of the primary AGB stars.

The standard model for Type II supernovae explosion, confirmed by the detection of the neutrinos emitted during the supernova explosion, predicts the formation of a compact object, usually assumed to be a neutron star. However, the lack of the detection of a neutron star or pulsar formed in the SN 1987A still remains an unsolved mystery. In this paper we suggest that the newly formed neutron star at the center of SN1987A may undergo a phase transition after the neutrino trapping time scale (~10 s). Consequently the compact remnant of SN 1987A may be a strange quark star, which has a softer equation of state than that of neutron star matter. Such a phase transition can induce the stellar collapse and result in a large amplitude stellar oscillations. We use a three dimensional Newtonian hydrodynamic code to study the time evolution of the temperature and density at the neutrinosphere. Extremely intense pulsating neutrino fluxes, with submillisecond period and with neutrino energy (> 30 MeV) can be emitted because the oscillations of the temperature and density are out of phase almost 180 degree. If this is true we predict that the current X-ray emission from the compact remnant of SN 1987A will be lower than 10^34 erg s-1, and it should be a thermal bremsstrahlung spectrum for a bare strange star with surface temperature of around ~10^7 K.

We use the Cambridge stellar evolution code STARS to model the evolution of 5-7 solar mass zero-metallicity stars. With enhanced resolution at the hydrogen and helium burning shell in the AGB phases, we are able to model the entire thermally pulsing asymptotic giant branch (TP-AGB) phase. The helium luminosities of the thermal pulses are significantly lower than in higher metallicity stars so there is no third dredge-up. The envelope is enriched in nitrogen by hot-bottom burning of carbon that was previously mixed in during second dredge-up. There is no s-process enrichment owing to the lack of third dredge up. The thermal pulses grow weaker as the core mass increases and they eventually cease. From then on the star enters a quiescent burning phase which lasts until carbon ignites at the centre of the star when the CO core mass is 1.36 solar mass. With such a high degeneracy and a core mass so close to the Chandrasekhar mass, we expect these stars to explode as type 1.5 supernovae, very similar to Type Ia supernovae but inside a hydrogen rich envelope.

We use the Cambridge stellar evolution code STARS to model the evolution and nucleosynthesis of zero-metallicity intermediate-mass stars. We investigate the effect of duplicity on the nucleosynthesis output of these systems and the potential abundances of the secondaries. The surfaces of zero-metallicity stars are enriched in CNO elements after second dredge up. During binary interaction, such as Roche lobe overflow or wind accretion, metals can be released from these stars and the secondaries enriched in CNO isotopes. We investigate the formation of the two most metal poor stars known, HE 0107-5240 and HE 1327-2326. The observed carbon and nitrogen abundances of HE 0107-5240 can be reproduced by accretion of material from the companion-enhanced wind of a seven solar star after second dredge-up, though oxygen and sodium are underproduced. We speculate that HE 1327-2326, which is richer in nitrogen and strontium, may similarly be formed by wind accretion in a later AGB phase after third dredge-up.