Protoplanetary disk lifetimes vs stellar mass and possible implications for giant planet populations
AI-generated Key Points
⚠The license of the paper does not allow us to build upon its content and the key points are generated using the paper metadata rather than the full article.
- The study investigates the relationship between protoplanetary disk evolution and stellar mass using a large sample of young stellar objects in nearby star-forming regions.
- The authors update the protoplanetary disk fractions presented in their previous work for 22 nearby associations aged between 1 and 100 Myr.
- A subsample of 1,428 spectroscopically confirmed members is used to examine the impact of stellar mass on protoplanetary disk evolution by dividing the sample into two stellar mass bins (2 M$_{\odot}$ boundary) and two age bins (3 Myr boundary).
- Objects are classified into three groups based on infrared excesses over photospheric emission: protoplanetary disks, evolved disks, and diskless.
- The authors find robust statistical evidence that protoplanetary disk evolution depends on stellar mass, with disks evolving faster or earlier around high-mass (>2 M$_{\odot}$) stars compared to low-mass stars.
- They also observe a roughly constant level of evolved disks throughout the entire age and stellar mass spectra.
- Such dependence could have important implications for gas giant planet formation and migration, which may contribute to explaining the apparent paucity of hot Jupiters around high-mass stars.
Authors: Álvaro Ribas, Hervé Bouy, Bruno Merín
Abstract: We study the dependence of protoplanetary disk evolution on stellar mass using a large sample of young stellar objects in nearby young star-forming regions. We update the protoplanetary disk fractions presented in our recent work (paper I of this series) derived for 22 nearby (< 500 pc) associations between 1 and 100 Myr. We use a subsample of 1 428 spectroscopically confirmed members to study the impact of stellar mass on protoplanetary disk evolution. We divide this sample into two stellar mass bins (2 M$_{\odot}$ boundary) and two age bins (3 Myr boundary), and use infrared excesses over the photospheric emission to classify objects in three groups: protoplanetary disks, evolved disks, and diskless. The homogeneous analysis and bias corrections allow for a statistically significant inter-comparison of the obtained results. We find robust statistical evidence of disk evolution dependence with stellar mass. Our results, combined with previous studies on disk evolution, confirm that protoplanetary disks evolve faster and/or earlier around high-mass (> 2 M$_{\odot}$) stars. We also find a roughly constant level of evolved disks throughout the whole age and stellar mass spectra. We conclude that protoplanetary disk evolution depends on stellar mass. Such a dependence could have important implications for gas giant planet formation and migration, and could contribute to explaining the apparent paucity of hot Jupiters around high-mass stars.
Ask questions about this paper to our AI assistant
You can also chat with multiple papers at once here.
⚠The license of the paper does not allow us to build upon its content and the AI assistant only knows about the paper metadata rather than the full article.
Assess the quality of the AI-generated content by voting
Score: 0
Why do we need votes?
Votes are used to determine whether we need to re-run our summarizing tools. If the count reaches -10, our tools can be restarted.
The previous summary was created more than a year ago and can be re-run (if necessary) by clicking on the Run button below.
⚠The license of this specific paper does not allow us to build upon its content and the summarizing tools will be run using the paper metadata rather than the full article. However, it still does a good job, and you can also try our tools on papers with more open licenses.
Similar papers summarized with our AI tools
Navigate through even more similar papers through a
tree representationLook for similar papers (in beta version)
By clicking on the button above, our algorithm will scan all papers in our database to find the closest based on the contents of the full papers and not just on metadata. Please note that it only works for papers that we have generated summaries for and you can rerun it from time to time to get a more accurate result while our database grows.
Disclaimer: The AI-based summarization tool and virtual assistant provided on this website may not always provide accurate and complete summaries or responses. We encourage you to carefully review and evaluate the generated content to ensure its quality and relevance to your needs.