Subscribe to the gold package and get unlimited access to Shamra Academy
Register a new userHow do binaries affect the derived dynamical mass of a star cluster?
137
0
0.0
(
0
)
Added by
M. B. N. Kouwenhoven
Publication date
2008
fields
Physics
and research's language is
English
Ask ChatGPT about the research
No Arabic abstract
The dynamical mass of a star cluster can be derived from the virial theorem, using the measured half-mass radius and line-of-sight velocity dispersion of the cluster. However, this dynamical mass may be a significant overestimation of the cluster mass if the contribution of the binary orbital motion is not taken into account. In these proceedings we describe the mass overestimation as a function of cluster properties and binary population properties, and briefly touch the issue of selection effects. We find that for clusters with a measured velocity dispersion of sigma > 10 km/s the presence of binaries does not affect the dynamical mass significantly. For clusters with sigma < 1 km/s (i.e., low-density clusters), the contribution of binaries to sigma is significant, and may result in a major dynamical mass overestimation. The presence of binaries may introduce a downward shift of Delta log(L/Mdyn) = 0.05-0.4 in the log(L/Mdyn) vs. age diagram.
rate research
Read More
Wave dark matter ($psi$DM) predicts a compact soliton core and a granular halo in every galaxy. This work presents the first simulation study of an elliptical galaxy by including both stars and $psi$DM, focusing on the systematic changes of the central soliton and halo granules. With the addition of stars in the inner halo, we find the soliton core consistently becomes more prominent by absorbing mass from the host halo than that without stars, and the halo granules become non-isothermal, hotter in the inner halo and cooler in the outer halo, as opposed to the isothermal halo in pure $psi$DM cosmological simulations. Moreover, the composite (star+$psi$DM) mass density is found to follow a $r^{-2}$ isothermal profile near the half-light radius in most cases. Most striking is the velocity dispersion of halo stars that increases rapidly toward the galactic center by a factor of at least 2 inside the half-light radius caused by the deepened soliton gravitational potential, a result that compares favorably with observations of elliptical galaxies and bulges in spiral galaxies. However in some rare situations we find a phase segregation turning a compact distribution of stars into two distinct populations with high and very low velocity dispersions; while the high-velocity component mostly resides in the halo, the very low-velocity component is bound to the interior of the soliton core, resembling stars in faint dwarf spheroidal galaxies.
The quenching rate is known to depend on galaxy stellar mass and environment, however, possible dependences on the hosting halo properties, such as mass, richness, and dynamical status, are still debated. The determination of these dependences is hampered by systematics, induced by noisy estimates of cluster mass or by the lack of control on galaxy stellar mass, which may mask existing trends or introduce fake trends. We studied a sample of local clusters (20 with 0.02<z<0.1 and log(M200/Msun)>14), selected independent of the galaxy properties under study, having homogeneous optical photometry and X-ray estimated properties. Using those top quality measurements of cluster mass, hence of cluster scale, richness, iron abundance, and cooling time/presence of a cool-core, we study the simultaneous dependence of quenching on these cluster properties on galaxy stellar mass M and normalised cluster-centric distance r/r200. We found that the quenching rate can be completely described by two variables only, galaxy stellar mass and normalised cluster-centric distance, and is independent of halo properties (mass, richness, iron abundance, presence of a cool-core, and central cooling time). These halo properties change, in most cases, by less than 3% the probability that a galaxy is quenched, once the mass-size (M200-r200) scaling relation is accounted for through cluster-centric distance normalisation.
The total mass of a distant star cluster is often derived from the virial theorem, using line-of-sight velocity dispersion measurements and half-light radii, under the implicit assumption that all stars are single (although it is known that most stars form part of binary systems). The components of binary stars exhibit orbital motion, which increases the measured velocity dispersion, resulting in a dynamical mass overestimation. In this article we quantify the effect of neglecting the binary population on the derivation of the dynamical mass of a star cluster. We find that the presence of binaries plays an important role for clusters with total mass M < 10^5 Msun; the dynamical mass can be significantly overestimated (by a factor of two or more). For the more massive clusters, with Mcl > 10^5 Msun, binaries do not affect the dynamical mass estimation significantly, provided that the cluster is significantly compact (half-mass radius < 5 pc).
Learning rate, batch size and momentum are three important hyperparameters in the SGD algorithm. It is known from the work of Jastrzebski et al. arXiv:1711.04623 that large batch size training of neural networks yields models which do not generalize well. Yao et al. arXiv:1802.08241 observe that large batch training yields models that have poor adversarial robustness. In the same paper, the authors train models with different batch sizes and compute the eigenvalues of the Hessian of loss function. They observe that as the batch size increases, the dominant eigenvalues of the Hessian become larger. They also show that both adversarial training and small-batch training leads to a drop in the dominant eigenvalues of the Hessian or lowering its spectrum. They combine adversarial training and second order information to come up with a new large-batch training algorithm and obtain robust models with good generalization. In this paper, we empirically observe the effect of the SGD hyperparameters on the accuracy and adversarial robustness of networks trained with unperturbed samples. Jastrzebski et al. considered training models with a fixed learning rate to batch size ratio. They observed that higher the ratio, better is the generalization. We observe that networks trained with constant learning rate to batch size ratio, as proposed in Jastrzebski et al., yield models which generalize well and also have almost constant adversarial robustness, independent of the batch size. We observe that momentum is more effective with varying batch sizes and a fixed learning rate than with constant learning rate to batch size ratio based SGD training.
Accurate chemical abundance measurements of X-ray emitting atmospheres pervading massive galaxies, galaxy groups, and clusters provide essential information on the star formation and chemical enrichment histories of these large scale structures. Although the collisionally ionised nature of the intracluster medium (ICM) makes these abundance measurements relatively easy to derive, underlying spectral models can rely on different atomic codes, which brings additional uncertainties on the inferred abundances. Here, we provide a simple, yet comprehensive comparison between the codes SPEXACT v3.0.5 (cie model) and AtomDB v3.0.9 (vapec model) in the case of moderate, CCD-like resolution spectroscopy. We show that, in cool plasmas ($kT lesssim 2$ keV), systematic differences up to $sim$20% for the Fe abundance and $sim$45% for the O/Fe, Mg/Fe, Si/Fe, and S/Fe ratios may still occur. Importantly, these discrepancies are also found to be instrument-dependent, at least for the absolute Fe abundance. Future improvements in these two codes will be necessary to better address questions on the ICM enrichment.
Log in to be able to interact and post comments
comments
Fetching comments
Sign in to be able to follow your search criteria
